Datasets:

OpenScholar

/

OpenScholar-DataStore-V3

like 5

Follow

OpenScholar 18

Socioeconomic inequalities in life expectancy and disability-free life expectancy among Chilean older adults: evidence from a longitudinal study Background Chile has one of the longest life expectancies of Latin America. The country is characterised by an important macroeconomic growth and persisting socioeconomic inequalities. This study analyses socioeconomic differences in life expectancy (LE) and disability-free life expectancy (DFLE) among Chilean older people. Methods The sample of the Social Protection Survey, a longitudinal study, was analysed. Five waves, from 2004 to 2016, were considered. The indicator was disability, defined as having difficulties to perform at least one basic activity of daily living. Type of health insurance was used to determine socioeconomic position (SEP). Total LE and DFLE were estimated with multistate life table models. Results At age 60, men in the higher SEP could expect to live 3.7 years longer (22.2; 95% CI 19.6–24.8) compared to men of the same age in the medium SEP (18.4; 95% CI 17.4–19.4), and 4.9 years longer than men of the same age in the lower SEP (17.3; 95% CI 16.4–18.2). They also had a DFLE (19.4; 95% CI 17.1–21.7) 4 (15.4; 95% CI 14.6–16.1) and 5.2 (14.2; 95% CI 13.4–14.9) years longer, compared to the same groups. Women aged 60 years in the higher SEP had a LE (27.2; 95% CI 23.7–30.8) 4.6 (22.7; 95% CI 21.9–23.5) and 5.6 (21.6; 20.6–22.6) years longer, compared to women in the medium and the lower SEP. The difference in DFLE, for the same age and groups was 4.9 and 6.1 years, respectively (high: 21.4;

232173147

95% CI 19.5–23.3; medium: 16.5; 95% CI 15.8–17.1; low: 15.3; 95% CI 14.6–16.0). Socioeconomic differences in LE and DFLE were observed among both sexes until advanced age. Discussion Socioeconomic inequalities in LE and DFLE were found among Chilean older men and women. Older people in the highest SEP live longer and healthier lives. Conclusion A reform to the Chilean health system should be considered, in order to guarantee timely access to care and benefits for older people who are not in the wealthiest group. Background Life expectancy (LE) in Latin America has been growing at a faster pace, compared to North America and Europe [1]. Chile is one of the countries with the highest LE at birth (80.3) of the región [2]. Ageing of the Chilean population has also been fast, with an ageing index of 87.2, considering an average of 51.8 for the región [2]. Apart from LE, it is important to know what proportion of those years will be lived in good health. Disability is among the most widely used indicators to assess health status of the older population, and to estimate disabilityfree life expectancy (DFLE) [3]. It has been reported that Chilean older women could expect to live more years disabled, compared to men of the same age [4,5]. There is evidence that along with the fast economic growth experienced by the country during the last decades, public policies addressing socioeconomic inequalities have been insufficient [6]. Furthermore, the organization of the health system, characterised by the coexistence of a public health insurance programme

232173147

and market-based private health insurance programmes, has deepened unequal access to health care [7]. Whereas in the public health system the employee contribution is based on income, in the private insurance system the contribution depends on individual health risk and number of dependents [7]. Private health insurance companies are allowed to create barriers to the affiliation of poorer and older people, by increasing premiums, deductibles and copayment, considering pre-existing conditions and health status [8]. Therefore, older people who are affiliated to the private health insurance system, are a selected group with a high income for the standard of the country, able to afford all these costs [8,9]. In 2017, 84.9% of the Chilean population aged 60 or more years was affiliated to the public health insurance system [10]. Previous studies have used the type of health insurance as a socioeconomic indicator, finding important differences in health and functional status among Chilean older people, depending on their socioeconomic position [11,12]. Recent analyses have reported marked socioeconomic inequalities in LE at old age in Chile [13,14]. To date, LE in a specific health status by socioeconomic position among Chilean older people has not been analysed. The aim of this study was to analyse socioeconomic differences in life expectancy and disabilityfree life expectancy among Chilean older people. Sample This study was based on secondary analyses of data from the Social Protection Survey (SPS). The SPS is aimed to collect information about social security in Chile, which includes a brief section about health status [15]. It is designed as a

232173147

fixed panel plus births longitudinal study [15]. The baseline sample in 2002 was representative of people affiliated to the Chilean pension system. In 2004, people not affiliated to the pension system and a refreshment sample were included, and the sample became nationally representative of the Chilean population aged 18 years or more [16]. The sample was recruited using a multi-stage stratified cluster sampling, considering the smallest territorial divisions as the clusters. The sampling method used to expand the sample in 2004 was similar. Since the sample from 2004 was nationally representative and disability started to be measured that year, this wave was the baseline measurement in our analyses, including people aged 60 or more years. A total of 3286 people had data on disability, type of health insurance and at least one follow-up measurement. Four succesive waves were considered (2006,2009,2012,2016). The data collection of the last wave of the SPS finished in July 2016 [17]. In the context of the present study, no ethical approval was required, since secondary analyses of anonymised and publicly available databases were performed. Outcome measures To estimate DFLE, disability was ascertained at each wave by asking participants if they usually needed help or had difficulties to perform basic activities of daily living, including bathing, dressing, eating and getting out of bed. A person that reported difficulties with at least one of these activities was classified as having a disability. Mortality was ascertained via retrospective measurement during the household interview. The information on dates of death collected in the SPS was validated

232173147

using official administrative data on the pension system history of active members, pensioned and deceased to July 2016. Socioeconomic indicator Health insurance type was used as socioeconomic indicator. People affiliated to the private health insurance system were considered as in high socioeconomic position (SEP). The medium socioeconomic position included people affiliated to the public health system who had received or still received a salary during their working life or contributed to the system to be affiliated (groups B, C and D). People who were affiliated to the system in group A (without income, hence not able to contribute), were classified as in low socioeconomic position. Statistical analyses Baseline characteristics of the sample by sex and SEP were reported. Chi square test and one-way ANOVA were used to determine SEP differences. The observation period was 15.7 years (from November 2004 to July 2016). All participants that had at least one follow-up measure were included in the analysis (n = 3286). A total of 13,280 records were available. A three state model was defined. State 1 was 'healthy' (without disability), state 2 was 'unhealthy' (with disability), and state 3 was 'dead'. Since 'dead' was an absorbing state, there were four possible transitions between these states, namely: healthy to unhealthy, unhealthy to healthy, healthy to dead and unhealthy to dead. When the state between two known states was missing, interval censoring was used. Right censoring was used when the health state at the end of follow-up was unknown, but alive. MSM for R [18] was used to fit multi-state

232173147

Markov models, in order to estimate the different expected health transitions, under the assumption of future evolution depending on the current state. A transition intensity matrix, Q, was calculated with all the observations, reflecting the instantaneous risk of moving from one state to another. Age (time-varying), sex and SEP were defined as covariates. To estimate total and marginal LEs, the ELECT (Estimating Life Expectancies in Continuous Time) package for R [19] was used. In order to do so, ELECT fits multinomial logistic regression models for state prevalence. Total, healthy and unhealthy LEs for men and women at age 60, 70 and 80, in each SEP, were estimated separately. To determine differences in these estimates, 95% confidence intervals were calculated. R version 4.0.3 was used for statistical analyses. Table 1 presents baseline characteristics for men and women in different SEP. A greater proportion of men and women in the higher SEP were in the youngest age group, and they were less likely to be in the oldest age group. Disability prevalence increased as SEP decreased. By the end of the follow-up, 34.1% (1119) participants had died. With respect to the number of deaths within groups, 41 (16.3%) participants in the higher SEP, 715 (34.5%) in the medium SEP, and 363 (37.8%) in the lower SEP, died during the observation period. The number of incident disability cases was 1371, and recovery from disability occurred 689 times. As observed in Table 2, in the case of men at age 60, those in the highest SEP expected to live 3.8

232173147

and 4.9 more years, compared to men in the medium and the lower SEP, respectively. These differences were present until 80 years of age, with a 2.3 and 2.9 years longer LE among me in the lower SEP, compared to men in the other SEP. At 60 years of age, women in the highest SEP expected to live 4.6 more years than women in the medium SEP, and 5.6 more years, compared to women in the lower SEP. Men aged 60 in the highest SEP expected to live 4 and 5.2 more years free of disability, compared to men of the same age in the medium and the lowest SEP. With respect to DFLE at age 60, women in the highest SEP had an advantage of 4.9 and 6.1 years, compared to women in the medium and lower SEP, respectively. Results Differences in DFLE between people in the highest SEP and people in the medium and the lower SEP were observed until 80 years of age. At this age, these differences reached 2.2 and 2.8 more years free of disability for men in the highest SEP, compared to the other SEP, and 2,8 and 3,4 more years for women in the highest SEP, compared to women in the other SEP. Although the absolute difference in years free of disability decreased with age, the difference in the proportion of years to be lived free of disability increased at older ages (Fig. 1). Although women in the medium and the lowest SEP had a longer total LE compared

232173147

to men of the same SEP, their LE free of disability was not longer. Differences in LE and DFLE were observed among men and women. Discussion Our results confirm that the trajectories of health, disability and mortality are more adverse for Chilean older people who are not in the highest SEP. We found significant differences in LE, depending on SEP, among Chilean older men and women. Comparing the highest and the lowest SEP, at 60 years the difference was 4.9 years for men, and 5.6 years for women, in favour of the wealthiest group. These differences are marked, but more conservative than those found in previous studies about the Chilean population [13,14]. This could be due to the fact that the data analysed in previous studies allowed them to make finer comparisons, such as the difference between the first and the tenth decile [14]. In that case, the difference between the poorest and the richest decile reached 7.7 years for men aged 60, and 17.8 years for women of the same age. People in the better-off group also expected to live more years free of disability, compared to people in the medium and lowest SEP. Since our indicator of disability was having difficulties or needing assistance to carry out basic activities of daily living, our results indicate that Chilean older people in the medium and the lowest SEP can expect to live more years of dependency. Less years free of disability have an impact on health, well-being and finances at a family level [20,21] and are

232173147

associated with greater demand for social and health care [22]. As observed in a previous study comparing population from England and the United States [23], the absolute difference between LE and DFLE by SEP decreased at older ages. However, in our study the difference in the proportion of years to be lived with disabilities, between the highest and the lower SEP, increased with age. In Chile, the type of health insurance is an expression of socioeconomic segregation among Chilean older people, particularly between the richest group and the rest of the population. Although the socioeconomic inequalities in LE and DFLE observed are associated with multiple factors through the life-course, the organisation of the Chilean health system and health insurance scheme themselves are a source of health inequalities, as previous studies have discussed [7,8]. Those who have access to private health insurance are not only richer, but also healthier, whereas the older population affiliated to the public health insurance are more likely to have more health problems and risks [24]. In 2004, 7.5% of people aged 60 or more years were affiliated to the private health insurance [25]. Around that time, 70% of the people affiliated to the private health insurance were under 50 years [26]. Since the private health insurance premiums increase progressively with age and health risks [24], the proportion of older people affiliated to the private health insurance continues to decrease, as observed in our study and reported previously [27]. According to this, a proportion of people in our sample who were affiliated to

232173147

the private health insurance system at baseline, probably migrated to the public health insurance, due to a limited payment capacity, health problems or both. If so, we would be underestimating the differences between the wealthiest group of older people in Chilethose who are able to stay in the private health insurance system until a very old ageand the rest of the population. Even though the Chilean public health system guarantees free access to health care for an important number of conditions, people affiliated to the private health insurance have more access to specialised medical services, laboratory tests and surgery [7]. In practice, the public health system is underfunded and insufficiently equipped, resulting in long waiting lists for specialist treatment and surgery, affecting more acutely the older, the poor and those with chronic conditions [28]. As previous reports have stressed [6,7,29], it is urgent to develop public policies able to meet the social and health needs of an increasingly older population in Chile. Previous studies have analysed two models that could explain social inequalities in health. One of them is social causation, which considers that socioeconomic position determines health during the life-course, whereas health selection considers that social mobility depends on the health status [30]. We did not have historical data that allow us to analyse health trajectories and social mobility during the life-course, hence we could not measure and cannot rule out the effect of health status on socioeconomic position. However, there are several elements that support the hypothesis that socioeconomic position is a determinant of

232173147

health inequalities in life expectancy and disability-free life expectancy among older people in Chile. First, European research suggests that the social causation has a higher importance to explain health inequalities in old age [30,31]. Second, Chile is one of the most unequal countries in the world [32], with a pattern of high concentration in a reduced group at the top and strong vertical barriers to social mobility [33]. A strength of this study is that it analysed longitudinal data and used multistate models, in order to take into account health transitions and time spent in each health state. For the first time, socioeconomic inequalities in DFLE among Chilean older people have been explored. There are some limitations to be considered. Although the sample of the SPS was nationally representative of Chilean people aged 18 years or more, people with missing data and those who have only one observation could not be included in the analyses. This could have affected representativeness and should be taken into account when interpreting the results. In fact, according to national data, in 2004 LE for men aged 60 was 19.7 years, and for women of the same age, 23.9 years. LE estimated with our data were 18.8 (95% CI 18.3-19.3) years for men, and 23.0 (95% CI 22.2-23.8) years for women. Also, the sample size affected the precision of our estimates. Therefore, the observed differences between socioeconomic positions could be smaller or larger than expressed by point estimates. Since a small proportion of older people are affiliated to the private health

232173147

insurance, the sample size for this group was the smallest, resulting in less precise estimates. Nevertheless, the differences found in LE and DFLE between this SEP and the other groups were significant. As mentioned above, we were not able to take into account migrations from the private to the public health insurance system as age increased, which might have led to an underestimation of differences in LE and DFLE between the higher and the other SEP. Also, the indicator used to determine SEP does not allow to make direct comparisons with international studies. However, type of health insurance is a good indicator of SEP among Chilean older people, since it is associated with education and income [11,34]. Conclusion Socioeconomic inequalities in LE and DFLE were found among Chilean older men and women. Older people in the highest SEP live longer and healthier lives. Our results support previous analyses that highlight the insufficient capacity of public policies in Chile to meet the social and health needs of an increasingly older population. A reform to the Chilean health system should be considered, in order to guarantee timely access to care and benefits for older people who are not in the wealthiest group.

232173147

Energy security and competition over energy resources in Iran and Caucasus region Energy security as a dominant factor in international stability is of great importance for major economies. The global energy market with its current level of supply and demand relies on energy sources in the Middle East, Caucasus, Central Asia and Russia. After the Fukushima disaster nuclear powers in Europe view renewable energy sources as a serious alternative. Europe’s energy vulnerability has deteriorated due to the Russia-Ukraine conflict. However, renewable energy sources are not large enough to replace nuclear power completely. This trend will continue with climbing demand especially in the natural gas sector as clean energy. In this research, Caucasus and Iran are considered the main sources and routes for energy transmission to the global market, including Europe. Caucasus plays a key role in bridging Europe and Asia. Also, Iran is an alternative for energy transmission to Europe after lifted sanctions. As part of the European active supply diversification policy Iran has capacity to reduce Europe’s energy dependency on Russia. However, changes in US new administration America First Policy is harmful for the EUs energy security. Caucasus aims to catch a large share of the European energy market since the Baku-Tbilisi-Ceyhan pipeline has started operations and Iran is also trying to expand its market to become a sustainable source of energy for major consumers. Therefore, Iran and Caucasus are considered reliable energy suppliers for Europe. In this regard, we analyze the best motivation for changing the direction new suppliers’ energy policies towards Europe and

41438679

suggest alternative solutions to compete with rival countries in order to enhance energy security. Introduction Energy, as a fundamental component of modern society and life, has a direct impact on human activity and it plays a critical role in socioeconomic development.Energy is deeply embedded in each component of mankind's development [1][2][3][4].Reliable and sustainable energy supply is therefore a fundamental condition for economic development and growth.It has also become essential for the smooth functioning of modern economies [5]. Increasing global energy demand, concerns about energy security such as availability of fossil fuels and the dependency on them, anthropogenic emissions of greenhouse gases and environmental degradation caused by energy generation from fossil fuels has stimulated debates about the future efficacy of fossil fuels [6][7][8][9].A legal definition of energy security is still in the process of being formulated and others [10] have contributed to this definition by stating it as a condition in which a nation and all, or most of its citizens and businesses have access to sufficient energy resources at reasonable prices for the foreseeable future free from serious risks of a major disruption of services. There are two new issues which put secure energy supply into question.First, global climate change has already started and its catastrophic consequences are increasing.The main reason for this is the growing consumption of fossil fuels.Second, the profound transformation of the energy market-the consequences of severe growth in demand, the state-monopolistic approach of countries with raw materials, shareholder value oriented strategies of private energy companies and the looming exhaustion of oil reserves-has

41438679

given rise to a situation in which satisfying demand is a matter of constant anxiety.On the other hand, the remaining oil and gas reserves are concentrated in politically unstable regions.The renewed focus on energy security is driven because of exceedingly tight oil markets and high oil prices.But it is also fueled by the threat of terrorism, instability in some exporting countries, geopolitical rivalries and countries' fundamental need for energy to increase their economic growth. However, the uneven distribution of energy supplies among countries has led to significant vulnerabilities [5].According to the International Energy Agency, 95% of the global economy is affected by the decisions of 5 or 6 states in the Middle East which are facing ethnic problems, political crises, terrorism, corruption and authoritarianism [11].Among others, Saudi Arabia and Iran play a major role in the energy security of the world [12].With this background there is renewed anxiety over whether there will be sufficient resources to meet the world's energy requirements in the decades ahead. Energy security as a dominant factor of international stability is of great importance for major economies.Changes in energy policy in Europe in the aftermath of the Fukushima disaster and Europe's deteriorating energy vulnerability situation due to the Russia-Ukraine conflict have led to continuous increase in demand especially in the natural gas sector as clean energy.In this research, Caucasus and Iran are considered the main sources and routes for energy transmission to the global market, including Europe.Caucasus aims to catch a large share of the European energy market through its Baku-Tbilisi-Ceyhan pipeline.Iran

41438679

is also trying to expand its energy supply market.Therefore, Iran and Caucasus are considered reliable energy suppliers for Europe.This research analyzes the motivation for changing the direction of the new suppliers' energy policies towards Europe and suggests ways to compete with rival suppliers in order to enhance market shares and improve general energy security.Examining Joint Comprehensive Plan of Action (JCPOA) and its possible effect on regional energy security and computation of energy security index for selected economies in EU for the period 1980-2014 could be considered as our contribution compared to previous literature.Due to conflict between Iran and Western powers, previous researchers didn't consider Iran as an important player in the region and its role in the regional energy security.Also, there is no analysis regarding the effect of JCPOA implementation on European energy security. The rest of this study is organized as follows.Section 2 discusses the importance of Iran in energy security.Section 3 gives details of energy supply and demand.The issues of energy security as a driver of alternative energy sources with a focus on the Iran and Caucasus region are discussed in Section 4 while Section 5 looks at energy resources.Section 6 discusses Iran's isolation and its implication for regional security.The final section provides a conclusion. Iran's Importance in Energy Security This section focusses on literature which has studied Iran's importance in the energy security of the world.Iran's potential contribution to Europe's energy security with emphasis on energy cooperation between Tehran and Brussels is assessed [13].According to his study progress in energy cooperation between Iran

41438679

and the European Union (EU) is likely to remain conditioned on reaching a compromise on allegations regarding Iran's nuclear program.Europe's reluctance to invest in Iran's hydrocarbon sector has provided great commercial opportunities for companies from other countries to make lucrative profits.According to the author, European (and American) technology is needed to fully utilize Iran's oil and gas deposits.Therefore, the development of these reserves will contribute to European and global energy security.This will also facilitate Europe's efforts to diversify suppliers and thereby reduce its heavy dependence on Russia. Iran's role in energy security at the regional and global levels is studied [14].According to her, under pressure from the United States which aims at the global isolation of Iran, the country has turned its eyes to Europe and China to meet its energy demands and has become a fruitful partner for Russia.Several factors such as neglecting Iran's legal rights in the Caspian Sea and lack of attention to the economic and political advantages of energy transmission through the country favor Russian policies.According to this study it is obvious that due to an idealist approach in its foreign policy Iran has lost the opportunity to play an effective and constructive role in energy security at the regional and global levels.The study suggests that Iran's unique geographical location between the Caspian Sea and the Persian Gulf is still usable and will serve Iran diplomatically.Also, despite the fact that opportunities were not seized, the ongoing regional and global developments continue to bring new opportunities.These opportunities can be exploited by realistically evaluating

41438679

the situation on the basis of national interests which benefit supplier and consumer nations. There is a need for infrastructure which will make it possible to supply an unlimited amount of energy from the South Caspian/Persian Gulf region to Europe, which is not completely available [15].Why isn't there a major pipeline from the South Caspian region to Europe which could be fed from all South Caspian and Persian Gulf states?This study gives political issues as the main reason for this.Private investors will not risk building a pipeline which might be disputed due to US sanctions against Iran.Also America's attitude towards Iran is especially frigid because of Iran's nuclear project and because Iran is allegedly destabilizing the situation in Iraq and beyond.The United States prevents Tehran and individual European investors' plans to transit Iranian energy resources via Turkey to the EU.However, Russia, alongside with the Caspian Sea Region gas suppliers and the Middle East account for about 90% of total global natural gas reservoir [16].Therefore, the critical role of Iran and Caucasus Region cannot be ignored by Europe. Energy Supply and Demand in Europe In the last two decades Europe's dependency on natural gas has significantly increased as natural gas has become a main source of energy.This dependency on natural gas emerged as a result of the EU's intergovernmental restrictions on CO 2 emissions, high emissions from coal-based generators and various obstacles in the rapid development of renewable energy [17].In 2009, about 26% of the primary energy consumption of all 27 EU member states was of natural

41438679

gas.With an annual growth rate of 2.7%, the total gas demand is projected to rise to 43% of primary energy consumption by 2030 [17]. Possible alternative energy sources for Europe Currently, about 50% of the European natural gas is imported by pipeline from outside the continent where the dominant supplier is Russia [16].Also, 32% of traded crude oil in Europe was sourced from Russia from 2015.European policymakers are trying to reduce their dependency on energy imports by enhancing renewable energy, developing energy efficiency and energy technologies.It is forecast that renewable energy production in Europe and Eurasia will increase to 330.7 Mtoe in 2035 compared to 115.5 Mtoe in 2013 [18].However, this trend is declining for oil production as estimates show that oil production levels will decrease to 803.3 Mtoe in 2035 as compared to 834.8 Mtoe in 2013 [19]. EU receives energy supplies from a number of countries around the world such as Norway, Russian, Central Asia, Caucasus and the Organization of Petroleum Exporting Countries (OPEC).EU works actively with these countries to get the best deals possible to increasingly diversify its energy sources and to prevent disruptions in supply.In future, additional energy demand from China and India should be met in competition by these sources as well. Norway is the world's third largest exporter of oil and gas after Saudi Arabia and Russia.In 2012, it accounted for about 31% of all EU's natural gas imports and 11% of its crude oil imports.Another energy source for EU is Russia as more than 40% of Europe's imported natural

41438679

gas through pipelines comes from Russia [18].Ukraine and Belarus' gas supply dependency on Russia is 74% and 100% respectively. A number of individual EU countries are heavily dependent on Russia particularly for natural gas (European Commission) 1 .This dependency leaves them vulnerable to supply disruptions, whether caused by political or commercial disputes and infrastructure failures.For instance, a 2009 gas dispute between Russia and the transit country of Ukraine, left many EU countries with severe shortages 2 .The Russia-Ukraine gas crisis not only illustrated to European states how strong their dependency is on Russian gas, but also emphasized the need to diversify their main energy supply sources as the flow of gas from Moscow can be unstable [20]. Importance of Iran and Caucasus for Europe energy security Iran has the highest natural gas reserves in the world (34 trillion cubic meters of natural gas).This amount is sufficient to satisfy current EU natural gas demand for 90 years.Despite this rich natural endowment, the country has not yet translated this potential into reality.Paradoxically, its natural gas production continues to be barely sufficient to satisfy its domestic demand.The Iranian natural gas industry most likely will remain focused on the domestic market, and on limited amounts of export within the region.In the aftermath of the nuclear deal, Iran is set to concentrate its energy strategy on the development of the oil sector.In this framework, more natural gas might be utilized for reinjection into oil fields in order to sustain growing oil production and exports.Iran will try to use its natural gas

41438679

resources to improve the competitiveness of its economy, through a larger share of power generation based on cheap natural gas, and through further investments in natural gas-fueled vehicles, in a move to reduce the domestic consumption of oil, which could thus be freed-up for exports 3 . Iran had a successful 2016 to capture a share of the European energy market.The bulk of international sanctions targeting Iran were lifted by JCPOA in January 2016 and the country's crude export to Europe has soared to 497,323 bbl/day during 2016 compared to 111,880 bbl/day during the sanction period (1st July 2012 to 31st December 2015).Table 1 shows the crude Iran's oil exported to Europe by destination.These figures could be considered as an indication for supply diversification made by European countries after JCPOA implementation on 16th January 2016. 1 https://ec.europa.eu/energy/en/topics/imports-and-secure-supplies/supplier-countries. 2 European Commission, https://ec.europa.eu/energy/en/topics/imports-and-secure-supplies/supplier-countries.Countries in Central Asia and Caucasus are rich in natural resources including oil and gas that can help EU diversify its energy supplies.Recognizing this potential EU has been participating in the development of their energy sectors.It has signed memorandums of understanding (MoUs) with Azerbaijan, Kazakhstan, Turkmenistan and Uzbekistan.These MoUs outline steps for further energy cooperation with these countries.The 12 countries that make up OPEC also provide EU with around 40% of its total crude oil imports.Of all OPEC countries, Saudi Arabia, Libya and Nigeria are the biggest individual suppliers, each having delivered over 8% of the EU's total oil imports in 2012.EU meets OPEC annually at the ministerial level to discuss a variety of issues including

41438679

the promotion of more stable oil prices and transparent markets. In response to energy security concerns, the European Commission (EC) released its Energy Security Strategy in May 2014.The strategy aims to ensure a stable and abundant supply of energy for European citizens and their economy (European Commission) 4 .The energy supply situation in Europe has deteriorated due to the Russia-Ukraine conflict.Therefore, Iran and Caucasus are considered reliable alternative energy suppliers for Europe.Figure 1 illustrates the importance of Iran and the Caucasus region as the main alternative sources of energy for Europe. Energy Security: A Driver for Alternative Energy Sources Energy security is affected by different factors including high oil prices, dependency on oil imports, depletion of oil fields, political instability in major oil exporting countries, disputes between transition countries and exporters and any disruption in energy supply.The level of security is estimated by the risk of supply disruptions or the costs incurred due to lack of security.The oil market is not considered a competitive market because more than 70% of the total crude oil reserves are located in OPEC [16].There are some geopolitical regions around the world which are energy sources.The most important ones include the Persian Gulf and the Caucasus region.There are many exporters in these regions which are sustainable sources of energy as they have huge sources of oil and gas.But some of them have suffered from political instability and military conflicts.Caucasus and Iran are considered as major sources of energy and alternative routes for energy transmission to the global market, especially to Europe.

41438679

Energy security in Europe's major economies Germany, Italy, UK and France are among the European countries which are members of G-7 and are considered major economics.They were also the largest importers of natural gas in Europe in 2014-2015.They imported 104.0, 50.2, 29.0 and 35.9 billion cubic meters (bcm) respectively of natural gas in 2015.Russia is the main source of natural gas transmission to Europe.We assessed energy security as a key factor of sustainable development for these countries by applying the Herfindahl-Hirschman Index (HHI) to measure the level of energy security.We calculated HHI of fuel-mix concentration for Germany, Italy, UK and France for 2000-2015.The results are given in Figure 2. Our findings show that the selected countries had HHI above 2000, indicating that these countries were highly concentrated in the use of fossil fuels as part of their primary energy consumption. We also derived a security index to measure the energy security index for some EU countries (see Table 2).This index was estimated by the World Bank based on net energy imports (as percentage of energy use) that are estimated as total energy use less domestic energy production [22].Table 2 shows that in particular Germany, Spain and Italy were highly dependent on importing energy.It should be noted that UK was a net exporter in 1990 and 2000, but this situation changed in 2010.High dependency on energy imports suggests that policymakers in EU have to find alternative sources of energy to replace Russia with more reliable supplier countries.Energy security or security of supply is related to three dimensions:

41438679

availability, affordability and sustainability.Iran and the Caucasus region have great potential to serve as alternative sources of energy for Europe.Considering the conflicts in Caucasus and the dispute between Iran and the western powers balanced energy security for all countries in the region is required.If there is lack of security for a country compared to the countries in its neighbourhood we cannot expect any cooperation.Any political instability, competition over energy resources or a dominant position in the region may lead to an unstable situation.If there is security cooperation, it should be applied to all parties.Otherwise an unbalanced situation will emerge.International cooperation from all parties is required to convert an Energy Security Complex to an Energy Security Community. Caucasus, strategic crossroads Caucasus is a valuable energy enhancing region due to its location at the crossroads between Europe and Asia.The region also possesses an important supply of energy [23].Integration of this region into the world market is significant for economic growth in these countries [24].In this context, Azerbaijan, Georgia and Turkey constructed energy transportation routes in Caucasus in the early 1990s to link these countries to Europe and Asia [25].In particular, the development of the Baku-Tbilisi-Ceyhan (BTC) oil pipeline and the Baku-Tbilisi-Erzurum (BTE) link Azerbaijan, Georgia and Turkey to Europe and Asia [26].It should be noted that Turkey and Georgia are only transit routes while Azerbaijan and the countries located on the east side of the Caspian Sea like Turkmenistan, Kazakhstan and Uzbekistan are energy suppliers. The BTC pipeline transits oil from Baku in Azerbaijan to Tbilisi in Georgia

41438679

and Ceyhan in Turkey.Its construction started in 1998 and the pipeline was completed in 2005 [27].The US had been pushing hard for BTC as the first pipeline specifically designed to export Caspian oil out of the region without going through Russia.It can transport 1 million barrels of oil a day from Azerbaijan via Georgia to the Turkish port of Ceyhan.Alongside BTC runs the BTE (or South Caucasus) gas pipeline through which Turkey imports gas from Azerbaijan.The BTE pipeline is used for transferring natural gas from Baku in Azerbaijan to Tbilisi in Georgia and further to Erzurum in Turkey.The construction of the BTE gas pipeline started in 2004 and was finished in 2007 [26]. The Caspian Pipeline Consortium (CPC) is a consortium and a major export pipeline to transport crude oil from the Kashagan and Karachaganak fields in Kazakhstan to the Novorossiysk on Russia's Black Sea coast.CPC will complete the expansion plan of this crude export route and raise the capacity to 67 mn t/yr by the end of 2017 from 52 mn t/yr [28].The CPC route is expected to transport 65.7 mn t/yr of crude this year of which more than 80% is sourced from Kazakhstan and the balance is supplied by the Russia.The most important parameter in the geopolitics of Caspian Region and Caucasus is control of oil and gas transition.Energy transportation through BTC, CPC or any possible pipeline would affect the region from different point of views including; energy perse (China), economic implications (Turkey, Iran, and the oil companies), a way to gain influence

41438679

and/or prevent others from doing so (the USA and Russia) [29]. Long-distance, cross-border pipelines are important to expand energy security and make an alternative to the many vulnerable checkpoints along the sea transportation routes [30].However, it is important to note that the construction of energy transportation routes has circumvented Armenia and Russia, which in turn, has shaped the economic, social and political integration in the Caucasus region [31].In recent years, independent of the central Government of Iraq, the Kurdistan region of the country has completed the construction of pipelines linking the region's oil and gas fields to Ceyhan in Turkey and Iranian energy supply routes with daily exports of more than 500,000 barrels of oil [32]. Iran, a geopolitical energy supplier Another alternative route for energy transmission to Europe with capacity to significantly reduce its energy dependency on Russia is Iran.The Russian Federation, which performs as a global player, is accused of taking advantage of the dependency of other countries by cutting off energy resource exports in order to reach its political aims.Also, the recent energy crisis between Russia and Ukraine has threatened European energy supply.Therefore, it has forced European countries to double their efforts to diversify their energy resources and energy transportation routes.Developments in the aftermath of the collapse of the Soviet Union along with the independence of its former republics in Caucasus and Central Asia have hastened the importance of Iran in the northern part of Hormuz Strait in the Persian Gulf [15]. On the other hand, Iran is located between two strategic energy

41438679

reserves of the world-Caspian Sea and the Persian Gulf, and it enjoys a unique situation.Besides Iran also borders Central Asian countries through its neighbor Turkmenistan and it has also become a fruitful regional partner for the Russian government in the post-Soviet Union era [33].According to an announcement by European politicians Iran's role in energy supply to Europe is becoming more important [15]. Iran is the world's third and OPEC's second largest oil exporter with a share of 15% in the Persian Gulf and 6% in the world oil market.Iran exports about 3,000,000 barrels of petroleum products per day, which is 1.3% of the total world exports.Iran exports liquefied petroleum gas (LPG) and distillate fuel oil.Iran is also the main importer of petroleum products in the Persian Gulf region.The country is ranked among the top countries that have many natural resources such as crude oil and natural gas.Its oil and gas are allocated 44% and 54% of the total energy share respectively [34].For over ten years Iran produced approximately 3.5 million barrels of oil per day (b/d) (currently 5% of the world production) with a rather narrow variance of less than 0.3 million b/d.Slightly more than 2 million b/d of production is available for exports.Iran is the fourth largest oil producer (behind Saudi Arabia, the US and Russia) and the fourth largest oil exporter (behind Saudi Arabia, Russia and Norway).Its reserves are larger than those of the US and Russia but smaller than those of Saudi Arabia and of Iraq [15].The oil and gas extracted from the

41438679

Kurdistan region is used mainly for domestic consumption, while that from the north and south is exported. Based on this discussion it is reasonable to state that Iran could be a candidate for changing EU's energy portfolio in a positive way.Iran with its large resources presents a new solution to the serious issue of security in the supply and sustainability of energy sources.This alternative seems promising as the current issues between Iran and the western powers are in the process of being solved through the recent agreement in Vienna on 14 July 2015.However, Iran's isolation and preventing international oil companies from investing in its oil and gas industry together with the country's reduction in the production of crude oil may hurt global energy security.In other words, the effect of the sanctions which have been imposed on Iran will influence the world energy market both in the short-and the long-term.The negative impact of these sanctions can be extended from a national to a global level in the future [35]. Competition Over Energy Resources by Superpowers Russia has had the will to dominate Caucasus and Central Asian countries since the collapse of the Soviet Union.There are some crucial parameters for this policy.First, Caucasus has huge sources of energy and it is considered a bridge between Russia and Europe.In fact, Georgia connects Caucasus to Europe and plays an important role in energy transmission to Europe.The challenges in Caucasus regional security after the Russia-Georgia conflict in August 2008 and explained some risks associated with the operating transit energy corridor in

41438679

southern Caucasus is studied [36].The conflict was considered a part of Russia's willingness to rebuild its former control sphere.The North Atlantic Treaty Organization (NATO) is interested in extending its borders to western neighboring countries of Russia including Georgia and Ukraine.By this argument energy resources and expansion of NATO borders could be seen as the main reason for the Russia-Georgia conflict.'Subduing Georgia would cut the west's vital energy connection (the Baku-Tbilisi-Çeyhan pipeline) to the Caspian Sea and to Central Asia' [37].Afterwards, Russia will be able to dominate energy flows from Azerbaijan. The proven oil reserves in various section of the Caspian Sea are different and are estimated at 298.4 billion barrels (Including Azerbaijan, Kazakhstan, Turkmenistan, Uzbekistan, Russia and Iran), corresponding to almost 17.5% of the total global proven reserves [16].But if we consider the Azerbaijan, Kazakhstan, Turkmenistan and Uzbekistan segment it has 38.2 billion bbl of reserves.These figures make one doubt that Caucasus and Central Asia are major energy resources as compared to the Persian Gulf.The case is different when we consider Iran in our analysis because of its huge energy sources and geopolitical situation.The total reserves in Iran are estimated at 157.8 billion bbl which is more than Russia at 102.4 billion bbl and four times higher than Caucasus and Central Asian countries together.Therefore, the size of reserves in Caucasus and Central Asian countries together with Russia is not comparable with Iran.The high cost of oil exploration in Caucasus may not be economically beneficial as much as it is in Iran though.But the importance of the

41438679

region in view of geopolitics, energy transmission and supply diversification cannot be insignificant. The Baku-Tbilisi-Ceyhan and Baku-Supsa oil pipelines along with the Baku-Tbilisi-Erzurum gas pipeline have promoted the importance of Caucasus as an energy corridor between Asia and Europe.Russia's invasion of Georgia and unilateral recognition of Abkhazia and South Ossetia made a significant change in the Caucasus from a geopolitical point of view.Russia wanted to dominate Georgia in order to cut off Azerbaijan and Central Asian countries and consequently it was able to enhance its energy monopoly in Europe significantly [36].Caucasus is like a magnetic field that Russia, Turkey and Iran are competing over with each other along with the United States and European Union.Russia was able to stop NATO's expansion into Caucasus by recognizing Abkhazia and South Ossetia [38]. Based on Henry Kissinger's thinking, capability and will are crucial variables in foreign affairs; although Russia was willing to dominate Caucasus countries but it did not have the capability [38].Caucasus and Central Asian countries are strong enough to not be dominated by Russia.In this regard, Russia is being challenged by China and the western powers in this region.The geopolitical situation in Caucasus is important for Turkey because it is the connection point between Europe and Asia.Therefore, Turkey is going to develop its influence in this region.At the same time, Turkey competes with Russia and Iran for energy transmission to catch a share of Europe's energy market.Although Turkey does not have any energy resources but its role as an energy corridor to the west combined with a

41438679

comprehensive investment program in refinery capacity are significant.The European Union relies on Russia, as it was the main source for 40% of its imported natural gas last year [16].Therefore, European countries are looking for supply diversification through alternative routes such as Caucasus and Iran to improve their energy security. Isolation of Iran and Its Implications for Regional Security Energy security as a dominant factor in international political stability is critical for major economics such as the United States, European Union, China and India.Any supply disruptions or fundamental changes may affect these economies through price volatility.Global energy markets have been affected by political instability and military conflicts in North Africa, the Middle East and Caucasus during the last decade.The Middle East is a formation of a multi-polar Regional Security Complex (RSC) that Iran plays as a regional power.Therefore, it is not possible to ignore it as an important factor for making and re-establishing a balance in the region.The Middle East and Caucasus faced imbalances after the Iran revolution in 1979 and RSC needs to be built in these regions through new policies.Sanctions imposed on Iran because of the country's nuclear activities targeted investments in oil, gas and petrochemicals, and exports of refined petroleum products.But in reality, Iran's political and economic isolation and the restrictions imposed on oil and gas field development can be considered a disruptive factor in global energy security. Iran might have been an alternative energy source for EU when it faced a challenge with Russia regarding its intervention in Ukraine.However, EU lost this opportunity

41438679

following Iran's isolation.After several rounds of nuclear talks between Iran and the western powers, UN sanctions were lifted on 16 January 2016.Consequently, Iran started increasing its oil production and exporting capacity in order to regain the market share that it had lost.Furthermore, Iran is going to introduce a new form of petroleum contracts in order to make them attractive for foreign investors.It is forecast that Iran will be a sustainable source of energy for the global market in the near future.Europe has the option to use the Iran-Turkey pipeline as an alternative for its natural gas transmission.But, Caucasus is another option available for exporting natural gas to Europe.In this regard, Caucasus has been the focus of attention of all parties including Europe, Russia, Turkey and Iran.Undoubtedly, Russia and Turkey are not interested in seeing enhanced regional importance for Iran. Considering the tensions between Russia and Ukraine, it is clear that Russia is not viewed as a reliable sustainable energy source for Europe.But, Iran and Caucasus have enough capability individually or combined to be an alternative channel for energy transmission to Europe.In this regard, foreign investments and technology transfers are required for developing the oil and gas fields in Iran and for constructing the necessary energy transfer infrastructure.Also, Caucasian countries suffer from lack of energy infrastructure.Energy security is important both from the point of view of supply and demand.Therefore, it will be a win-win situation if major energy consuming EU countries undertake investments in energy supplier countries to build infrastructure, transmission and upstream and downstream facilities. The

41438679

military superiority of the US and the lack of a rival superpower to create a balance has led to replacement of global political space by a much more monolithic domination by the west.In the absence of an ideological challenger within or adjacent to the core, the Western powers in general and the US in particular, can impose much more demanding legal, social, financial, and political conditions [29].Iran is considered as a regional power in Middle East and due to its critical role in views of energy security in the regional and global levels, its importance cannot be ignored.In order to convert an energy security complex to an energy security community, international cooperation among all parties is required.In our view, Western powers should take into account the fact that Iran can play a polar in regional stability and global energy security.Iran has a huge source of oil and gas and may be an effective alternative for Europe to achieve supply diversification in order to reduce its energy dependency to Russia.Therefore, international cooperation and performing JCPOA in case of Iran, may reduce the Iran-US conflict. However, conservative policymakers in the United States supported by Israel and rival Arab countries in the Middle East have come up with additional impediments in rejoining Iran to the international community.It is clear that Iran has the will and the capability to play its role as a regional power in the Middle East.The country's open interventions in the security and stability affairs of Iraq, Syria and Yemen are evidence of its desire and

41438679

capacity.Considering the geopolitics of Iran in the Middle East and its capabilities, it can be predicted that it will be a key factor in global energy security in the next decades as it has large energy reserves and a strategic position for energy transmission.Obviously, decisions made by Iranian policymakers and their interactions with western countries, particularly the United States, will be vital for this purpose. Conclusion There is a close relationship between diversifying the supply of energy sources and enhanced energy security.The Russia-Ukraine gas crisis not only illustrated to EU how strongly it is dependent on Russian gas, but also emphasized its need to diversify its energy supply sources.Iran and Caucasus are considered the main sources and routes for energy transmission to the global market, including to Europe.Security in Caucasus is affected by Russia's intervention in Georgia and the possibility of Armenia and Azerbaijan getting support from Russia, Turkey and Iran.Although NATO was interested in expanding its borders to include Caucasian countries but the Russia-Georgia conflict was an impediment in this.In the case of Iran, developing oil and gas fields, economic capacity and the country's international relations were significantly and negatively affected by the sanctions imposed against it by the United States and its European allies. Conservative policymakers in the US supported by Israel and regional rivals like Saudi Arabia are going to keep Iran isolated after the lifting of UN sanctions.However, the geopolitical situation of Iran in the Middle East and its role as a regional power make achieving this target difficult.Russia is trying to

41438679

rebuild the influence of the former Soviet Union but it does not have enough capacity to enforce this on states in Caucasus and Central Asia.The importance of energy transmission has converted Caucasus and Iran into a game of chess played by great powers like the US that have tried so as to bypass Iran and Russia by constructing the BTC pipeline.Considering Iran's geopolitics in the Middle East and its capabilities, it can be predicted that Iran will be a key factor in global energy security in the next decades due to its massive energy resources and strategic location for energy transmission.Caucasus too can be an alternative for natural gas transmission instead of Turkey.Therefore, the situation of Iran-Caucasus will be enhanced if this plan is realized in the future.International cooperation and performing JCPOA completely in case of Iran, may reduce parts of Iran-US conflict.Due to having a huge source of oil and gas, Iran has sufficient capability to serve as an effective alternative for Europe to achieve supply diversification in order to reduce its energy dependency on Russia.The combination of Iran resources and Caucasus as a corridor of energy supply could be a proper solution for enhancement of energy security in Europe.However, recent changes in US policy towards Iran is harmful for the EUs energy security. Table 1 . Crude oil exported to Europe by Iran after JCPOA (bbl/d).

41438679

A Novel Tactile Function Assessment Using a Miniature Tactile Stimulator Several methods for the measurement of tactile acuity have been devised previously, but unexpected nonspatial cues and intensive manual skill requirements compromise measurement accuracy. Therefore, we must urgently develop an automated, accurate, and noninvasive method for assessing tactile acuity. The present study develops a novel method applying a robotic tactile stimulator to automatically measure tactile acuity that comprises eye-opened, eye-closed training, and testing sessions. Healthy participants judge the orientation of a rotating grating ball presented on their index fingerpads in a two-alternative forced-choice task. A variable rotation speed of 5, 10, 40, or 160 mm/s was used for the tactile measurement at a variety of difficulties. All participants met the passing criteria for the training experiment. Performance in orientation identification, quantified by the proportion of trials with correct answers, differed across scanning directions, with the highest rotation speed (160 mm/s) having the worst performance. Accuracy did not differ between vertical and horizontal orientations. Our results demonstrated the utility of the pre-test training protocol and the functionality of the developed procedure for tactile acuity assessment. The novel protocol performed well when applied to the participants. Future studies will be conducted to apply this method to patients with impairment of light touch. Introduction Somatosensation, also known as the sense of touch, is a crucial and commonly used sense in everyday life. Somatosensory deficit is a common symptom in both peripheral and central neurological diseases and insults [1], such as lesions of sensory receptors, lesions of the peripheral nerves,

256718810

or impairments to haptic representation in the central nervous system (specifically in the primary sensory cortex, such as stroke) [2,3]. Therefore, tactile spatial acuity (TSA) is commonly measured to evaluate somatosensory function [4,5] and yields important information for diagnostic, functional, and prognostic purposes. Tactile spatial and directional sensitivity on a body part is mainly determined by two factors, the density of peripheral mechanoreceptors and their receptive field properties [6,7]. Several assessments have been developed to objectively measure tactile spatial resolution in various physiological or pathological conditions. Two-point discrimination, a tactile spatial discrimination task, is the most widely used for this purpose [6]. In recent decades, Sensors 2023, 23, 1844 2 of 13 the utility of two-point discrimination has come under scrutiny because it can only measure at the threshold of a just-noticeable difference instead of the limit of spatial resolution, and most importantly, it tends to yield inconsistent results because it cannot control nonspatial cues [5,6]. The JVP dome, a grating dome that assesses sensory capacity in grating orientation discrimination [2], is regarded as the standard method used to qualify the tactile threshold for the spatial resolution [3]. Grating orientation sensitivity is considered suitable for assessing spatial acuity because it is affected by the density of innervation and varies with the somatosensory function of the fingerpad [8]. Nevertheless, the use of the JVP dome requires considerable skill and the indentation depth on the skin cannot be accurately controlled because it is delivered by the examiner's hand, making it necessary to develop a fully automatic method that

256718810

allows the tactile orientation discrimination task to be performed. The tactile orientation discrimination task has been performed by healthy people [5,6], patients with neurological disorders [3,9,10], and blind Braille readers [11,12]. Studies have reported that the TSA of blind Braille readers is superior to that of healthy people [11,12], whereas the TSA of patients with neurological disorders is worse [3,9,10]. Moreover, TSA can improve with training [11,13]. The tactile orientation discrimination task has also been performed by children [14] and older adults [15,16], indicating that older age is correlated with a decline in tactile spatial resolution. The miniature tactile stimulator (MTS) was developed by our group for performing automatic tactile stimulations that can cover a variety of movement directions and speeds; these simulations are used to characterize the human ability to perceive shape and motion by touch [17]. MTS consists of three independently controlled micromotors, providing three degrees of freedom, which are the grating ball's direction of movement, speed of rotation, and depth of vertical indentation on the skin [17]. We successfully demonstrated the functionality of the MTS for measuring the performance of tactile motion discrimination in healthy participants [17][18][19], and the MTS thus offers unique potential as a fully automatic and standardized measurement of tactile acuity. In the present study, we developed a method where MTS is used for an orientation discrimination task; this method can be applied to the measurement of the severity of somatosensory deficits. Through MTS, motion stimuli were applied to the participant's fingerpad, and the participant reported the received orientation of

256718810

the moving grating ball. We describe how the measurement was performed during the training and testing phases. Specifically, the orientation discrimination task was performed at a variety of motion speeds to characterize how much motion speed affected tactile acuity. We demonstrated that MTSbased orientation discrimination tasks can be effectively applied to evaluate tactile acuity in healthy participants. 1. This study and the experimental protocol were reviewed and approved by the institutional review board (IRB) of the Chang Gung Medical Foundation (IRB no.: 202001628B0A3). All of the methods were performed in accordance with the regulations of the Taiwan Human Subjects Research Act and the guidelines of the Declaration of Helsinki 1975. The details of the study and the procedures were clearly explained to each participant. 2. Sixteen healthy participants were recruited for the main experiment, and written informed consent was obtained from all participants ( Figure 1). 3. The inclusion criteria were as follows: (1) aged between 40 and 65 years old, and (2) normal cognitive function as indicated on a judgment, orientation, memory, abstract thinking, and calculation (JOMAC) scale. Experimental Setup 5. The MTS system was set up on a desk and connected to a PC through RS-232 ports ( Figure 2A). 6. The MTS system comprised a three-motor controller for manipulating the indentation depth, directions of motion, and rotation speeds of the grating ball; a finger support surface for the participant to place their finger on; and a human-machine interface for controlling the MTS [17]. 7. MTS consisted of three degrees of freedom (DoF) (one

256718810

of which is illustrated in Figure 2B): rotation for producing motion, vertical excursion for controlling the depth of indentation into the skin, and arm orientation for controlling the direction of motion. Each DoF was controlled by a single motor: two DC motors controlled the rotational 5. The MTS system was set up on a desk and connected to a PC through RS-232 ports ( Figure 2A). 6. The MTS system comprised a three-motor controller for manipulating the indentation depth, directions of motion, and rotation speeds of the grating ball; a finger support surface for the participant to place their finger on; and a human-machine interface for controlling the MTS [17]. 7. MTS consisted of three degrees of freedom (DoF) (one of which is illustrated in Figure 2B): rotation for producing motion, vertical excursion for controlling the depth of indentation into the skin, and arm orientation for controlling the direction of motion. Each DoF was controlled by a single motor: two DC motors controlled the rotational motion and direction of motion, while one DC motor controlled the depth of indentation. 8. The user interface (UI) included several parts: the first part was for registering information on the participants' general characteristics; the second part was for setting up the three control parameters (indentation depth, motion direction, and rotation speed) of the grating ball; the third part was for recording the participant' responses in the orientation discrimination task, where they selected the corresponding vertically and horizontally oriented pictures on the screen; and the fourth part was for analyzing the

256718810

results of the task immediately after each training or testing session. 9. The grating ball, engraved with 2 mm sine-wave gratings with a peak-to-peak amplitude of 0.5 mm, was installed on the stimulator ( Figure 2C). 10. On each trial, the grating ball first rotated the orientation of the grating ball ( Figure 2B), and then rotated the grating ball to reach the defined surface scanning speed. Next, the rotating grating ball compressed the fingerpad with a fixed indentation depth of 1 mm for 1 s to present the tactile stimulation. Finally, the rotated grating ball moved upward, left the skin, and then stopped rotating. 11. Headphones and a blindfold were prepared, which prevented the participants from seeing the grating orientation and hearing the motor noise of the MTS. 12. Information on the participants' basic characteristics was registered, and tactile stimuli parameters were set into the control program via the UI. 13. The participants' index finger was secured inside the finger holder using Velcro. The forearm and hand were positioned palm side up on the forearm supporter and the base of the MTS. 14. The grating ball was adjusted to the surface of the participants' finger, to make sure that the indentation of the tactile stimuli during tests was precise ( Figure 2B). motion and direction of motion, while one DC motor controlled the depth of indentation. 8. The user interface (UI) included several parts: the first part was for registering information on the participants' general characteristics; the second part was for setting up the three

256718810

control parameters (indentation depth, motion direction, and rotation speed) of the grating ball; the third part was for recording the participant' responses in the orientation discrimination task, where they selected the corresponding vertically and horizontally oriented pictures on the screen; and the fourth part was for analyzing the results of the task immediately after each training or testing session. 9. The grating ball, engraved with 2 mm sine-wave gratings with a peak-to-peak amplitude of 0.5 mm, was installed on the stimulator ( Figure 2C). 10. On each trial, the grating ball first rotated the orientation of the grating ball ( Figure 2B), and then rotated the grating ball to reach the defined surface scanning speed. Next, the rotating grating ball compressed the fingerpad with a fixed indentation depth of 1 mm for 1 s to present the tactile stimulation. Finally, the rotated grating ball moved upward, left the skin, and then stopped rotating. 11. Headphones and a blindfold were prepared, which prevented the participants from seeing the grating orientation and hearing the motor noise of the MTS. 12. Information on the participants' basic characteristics was registered, and tactile stimuli parameters were set into the control program via the UI. 13. The participants' index finger was secured inside the finger holder using Velcro. The forearm and hand were positioned palm side up on the forearm supporter and the base of the MTS. 14. The grating ball was adjusted to the surface of the participants' finger, to make sure that the indentation of the tactile stimuli during tests

256718810

was precise ( Figure 2B). The participants were informed of the procedure, safety precautions, and purpose of the experiment. The participants were then asked to sign the written informed consent form. 16. The experimenter measured TSA using the MTS and comprised three sessions: eyeopened training, eye-closed training, and testing. 17. The MTS performed tactile stimulations on the index fingertip of the participant's right hand. The procedure was noninvasive and painless, and the total duration was approximately 30 min. 18. In the main experiment, in each trial, the participants were presented with one of the stimulus conditions from a factorial design of 16 combinations (4 directions × 4 speeds) in a pseudorandom order. Specifically, the directions were 0 • , 90 • , 180 • , or 270 • and the surface moving speeds were 5, 10, 40, or 160 mm/s. The tactile stimulus had an indentation depth of 1 mm, and each stimulus lasted for 1s. 19. The reference frame was defined with respect to the forearm when the patients were in a neutral position. Directions of 0 • and 180 • indicated that the grating ball was moving rightward and leftward, respectively, and thus yielded a vertical orientation as the grating orientation was orthogonal to the scanning direction. Similarly, 90 • and 270 • indicated that the grating ball was moving distally and proximately, respectively, and thus yielded a horizontal orientation ( Figure 2D). Although the participants were presented with one of the four directions, the participants' task was to report the perceived grating orientation

256718810

(horizontal or vertical). Eye-Opened Training Session 20. The participants' fingers were fixed at the appropriate location inside the finger holder (Please see numbers 9 and 10 in Section 2.2). 21. The aim of this task was to confirm that the participants understood the task structure. 22. The participants were instructed to visually inspect the orientation of the grating engraved on the stimulation ball during stimulation. 23. The tactile stimulation was identical to that described in numbers 4 and 5 in Section 2.2.1, with a fixed surface moving speed of 40 mm/s. We employed a total of four combinations (4 directions × 1 speed) and repeated each combination three times, yielding a total of 12 trials (4 directions × 1 speed × 3 repetitions). In each trial, the participants were thus presented with one of the vertical or horizontal orientations (see number 5 in Section 2.2.1). As a two-alternative forced choice design, the participants verbally reported the perceived orientation after each tactile stimulus. 24. Based on our inclusion criteria of cognitive function, we expected that all participants would meet the passing criteria during training. Eye-Closed Training Session 25. This was a 24-trial task where the participants' eyes remained closed. 26. The participants were asked to wear a blindfold and headphones that played white noise so that they could not see or hear the stimuli. 27. The aim of the eye-closed training session was to ensure that the participants could perform the task under conditions analogous to those of the testing session. 28. Each stimulation was performed as

256718810

described in number 4 and 5 of Section 2.2.1, with a fixed surface moving speed of 40 mm/s. We employed a total of four combinations (4 directions × 1 speed) and each of the combinations was repeated six times (4 directions × 1 speed × 6 repetitions = 24 trials). 29. The accuracy, quantified by the probability that the participants would give a correct answer, was computed from the aforementioned 24 trials. If the participants' accuracy was ≥75%, we proceeded to the testing session; if not, we repeated the eye-opened training (number 3 and 4 in Section 2.2.2) to make sure they understood the task. , for a total of 96 trials. 33. The results of each patient in the trial were analyzed. 34. In order to investigate the effect of motion of rotation with the grating edge on the fingerpad, some of the participants (8 of 16) also received additional stimulations by using 1 mm and 4 mm sine-wave grating balls ( Figure 2C). 35. The procedure for the stimulation using 1mm and 4mm grating balls was the same as that of the 2 mm grating ball. 36. The system usability scale (SUS) [20,21] questionnaire was applied to examine the usability of MTS in participants who had no experience operating the device. 37. The recruited participants were briefly instructed with our MTS-based protocol. 38. Our examiner firstly operated MTS to test the participants' tactile acuity using the protocol. 39. Next, the participants would try to operate MTS and test the examiner's tactile acuity independently. 40.

256718810

Finally, the participants completed the SUS questionnaire. Data Analysis 41. The statistical analyses were performed using Statistical Program for Social Sciences (SPSS). The Wilcoxon signed-rank test was applied to compare the inter-orientation accuracy between the vertical and horizontal orientations. The Friedman test was applied to compare the accuracy between speeds, and if the results were significant, the Wilcoxon signed-rank test was used as the post hoc test. The repeated measure ANOVA was applied to the full 3 wavelength × 4 speed model, in which factors are wavelength (1, 2, and 4 mm sine-wave grating ball) and speed (5, 10, 40, and 160 mm/s). The confidence interval was corrected using the Bonferroni method. 42. All data are presented in terms of the mean ± standard error of the mean (SEM) with 95% confidence intervals. 43. The SUS questionnaire consisted of 10 statements with five positive and five negative statements and each question had a five-point scale (strongly disagree = 1 point and strongly agree = 5 points). 44. For analysis of SUS, the odd-numbered questions, Q1, 3, 5, 7, and 9, were positive questions, and the recorded scores were: original score − 1. 45. Even numbered questions, Q2, 4, 6, 8, and 10, were negative questions, and the recorded scores were: 5 − original scores. 46. The recorded scores from the ten questions were summed up and then multiplied by 2.5 to yield the total SUS score. 47. The total SUS score ranged from 0 to 100. When it came to an acceptable SUS score, products were

256718810

at least passable when SUS scores were over 70. Good products could score in the mid-70s to low-80s. Excellent products would score better than 85. Products with scores less than 70 needed to be considered for scrutiny and improvement [22]. Results Sixteen healthy participants (eight men and eight women, aged 48.50 ± 1.50 years) were enrolled (Table 1). We tested their right-hand tactile accuracies, but two of them were excluded due to carpal tunnel syndrome in at least one of their hands. All of the remaining participants completed the training and formal experiments and no adverse events were reported. In the eye-opened training session, all of the participants achieved 100% accuracy (n = 14). Data from the training session indicated that the participants fully understood the experimental protocol. This supported the effectiveness of the instruction to the participants. In the eye-closed training session, eleven of the fourteen participants passed the threshold (threshold = 75%, Table 2). Although participants #3, #13, and #16 initially failed the eye-closed training (accuracy = 62.5%, 45.83%, and 58.33% respectively), these participants passed after receiving an additional eye-opened training session (accuracy = 100%). The demographic data in the testing session are shown in Table 3. The mean accuracy was 89.36% ± 1.12% (n = 14), which exceeded the threshold of 75%. The results also demonstrated that accuracy in the vertical (87.35 ± 1.55%) and horizontal (91.37 ± 1.49%) orientations did not significantly differ (p = 0.08, Figure 3A). Subsequently, we characterized the effect of motion speed on the performance of orientation identification.

256718810

The results demonstrated that tactile accuracy differed across speeds (p < 0.001, Friedman test), and the post hoc test revealed that accuracy at the highest speed (160 mm/s) was significantly inferior to that in each of the other three speeds (5, 10, and 40 mm/s, Figure 3B). This finding indicates that scanning speed might affect the participants' performance, which could be applied for assessing tactile acuity under various conditions. We examined whether the wavelength of the grating affects the accuracy of orientation discrimination, a phenomenon that could support that the percept was mainly mediated by spatial information presented by the grating rather than the shear force. To this end, 8 of the 14 participants were presented with sinusoid grating balls with wavelengths of 1, 2, or 4 mm. The results showed that accuracy was significantly modulated by wavelength (p = 0.006) and the post hoc analysis showed that the 1mm grating had a worse accuracy compared with the 4mm grating ( Figure 4A, p = 0.019). The performance was thus better with a wider wavelength of grating. Discussion MTS is a robotic tactile stimulator that could be applied for tactile function capability assessments [17,18,23]; it is novel because a standard and feasible method has yet to be established. In the present study, we developed an MTS-based program for tactile acuity evaluation. The results demonstrated the following: (1) the examining process could be easily performed by examiners, and all of the participants passed both training sessions with accuracy above the predefined threshold, a finding that supports the

256718810

utility of the In addition, we examined whether the scanning speed affected the accuracy of orientation discrimination. The results showed that accuracy was significantly modulated by motion speed (whole ANOVA model here p = 0.009) and the post hoc analysis showed that 160 mm/s had a worse accuracy compared with 5 and 160 mm/s ( Figure 4B, p = 0.042), indicating that the performance decreased at extremely high speeds. Moreover, there was no interaction effect between the wavelength and scanning speed ( Figure 4B, p = 0.733), indicating that the aforementioned wavelength effect was independent of the scanning speed. Finally, we evaluated the usability of the MTS protocol using the SUS questionnaire in seven naïve examiners (n = 7) ( Table 4). The results demonstrated that the overall SUS score was 88.57 ± 3.61, indicating that the usability of the present protocol was excellent (the overall SUS score >85) [22]. The recorded scores in all items were ≥3 in all questions, except item Q4 (Table 5), "I required technical assistance to use the miniature tactile stimulator", in which the score was only 2.71 ± 0.52, suggesting that the participants might need assistance when operating the device. In summary, our results demonstrated that the novel procedure for tactile stimulation could be applied as a psychophysical assessment for examining the tactile acuity of healthy individuals. Discussion MTS is a robotic tactile stimulator that could be applied for tactile function capability assessments [17,18,23]; it is novel because a standard and feasible method has yet to be established. In the

256718810

present study, we developed an MTS-based program for tactile acuity evaluation. The results demonstrated the following: (1) the examining process could be easily performed by examiners, and all of the participants passed both training sessions with accuracy above the predefined threshold, a finding that supports the utility of the pre-test training; (2) neither discomfort nor adverse events were reported by the participants, suggesting the acceptability of this designed tactile measurement protocol; (3) all participants had above-threshold accuracy (accuracy ≥ 75%) when identifying the directions of motion at lower speeds (5, 10, and 40 mm/s), indicating that most healthy participants could perform the present protocol; and (4) the accuracy of orientation identification decreased at higher speeds, suggesting that higher speed (160 mm/s) conditions could be more challenging for tactile testing. In summary, the present study developed an evaluation protocol that could apply MTS for the automatic testing of tactile acuity. This is the first protocol designed for the assessment of tactile function using an automatic robotic system with MTS. Because we aimed to apply this program to patients with somatosensory deficits, the protocol was designed with an adjustable degree of difficulty that can discriminate between various severities of somatosensory impairment. Considering user comfort in the testing program, we modified the body postures through which the participants received the assessment. In addition, the duration of the entire examination procedure was optimized and was shown to be tolerable by all participants. Tactile orientation discrimination tasks have been applied to distinct body parts, such as the lip, tongue, and fingers.

256718810

The levels of TSA sequenced were highest to lowest in the lip, tongue, index or ring finger, and little finger, respectively [2,4,24]. In the present study, we chose the index finger as our measuring target for its advantages of fine TSA and easy manipulation compared with the lip or tongue. An important question is whether the participant relied on the stretch force induced by the scanning movement rather than the spatial-temporal information presented by the grating orientation [19,25]. Indeed, in the present setup, the shear force was always orthogonal to the grating orientation, so that it was possible that the participants used the direction of shear force to infer the grating orientation. Using sine-wave grating with a variety of wavelengths, we observed a strong effect of wavelength on the accuracy, a finding that is reminiscence of the effect of grid width on static orientation judgement [5,26]. Furthermore, speed and wavelength are two independent factors that influence accuracy, suggesting that wavelength as a spatial factor has a distinct mechanism to affect performance and this mechanism is not affected by speed. To this end, these findings support that the participants' performance was mainly determined by spatial-temporal cues. For usability, the results showed that most of the SUS items had high scores, indicating that a naïve examiner could operate the MTS well after brief instruction and training. However, we found the lowest score in item Q4, suggesting that non-experienced operators might require help from others. For example, the operator sometimes hesitated to select the settings on the operational interface

256718810

software and needed assistance from the experienced operator. In our opinion, protocols that simplified and optimized operative steps are expected to improve the users' satisfaction and its clinical usability. The MTS system and present study have several limitations. First, the sample size was small because this was a pilot study for assessing the feasibility and acceptability of the MTS-based program for measuring tactile acuity. Second, it was difficult to apply this tactile function assessment instrument to some parts of the body, such as the lip and tongue, because more adjustment was required to apply the MTS on other body parts. Third, although the protocol was optimized in the present study, it still took up to 30 min for each participant to complete the whole process. Fourth, it is possible that the subjects still had minute finger movements when the fingers were placed on the finger holder during the experiment. Without continuous monitoring of finger position, it was hard to make sure the participants remained in their assigned position throughout the experiment. In addition, participants may have still been able to perceive the direction of motion, relying on the spatiotemporal cues provided by the miniature ball's rotation and stretch cues-the rotation exerted on the skin surface. Indeed, tactile direction discrimination is the ability to recognize the tactile motion direction of an object moving across the skin [27,28]. Sensing the direction of a tactile stimulus relies on spatiotemporal cues and skin stretch through the simultaneous processing of information signaled by large myelinated afferent nerve fibers and dorsal column

256718810

pathways [29,30]. Tactile direction judgments are sensitive to the direction of skin stretch, whereas low friction stimuli with minimal skin stretch, such as the rolling ball, provide only spatiotemporal cues (successive positions cues) for the direction of motion [31][32][33]. Furthermore, SA1 receptors are at least ten times more sensitive to moving than to static stimuli [25]. The MTS system is an automatic, noninvasive, and quantitative method for assessing tactile acuity. Our study demonstrated that this novel MTS-based protocol is safe, accessible, tolerable, and feasible as a simple standardized measurement of tactile acuity. We will conduct further studies where we will apply this program to evaluate somatosensory deficits in patients with peripheral neuropathies or central neurological disorders. Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. Data Availability Statement: The data that support the findings of this study are available from the corresponding author upon reasonable request.

256718810

Predicting the Porosity Formed in Freeze Casting by Artificial Neural Network Freeze casting has been increasingly applied to process various porous materials. A linear relationship between the final porosity and the initial solid material fraction in the suspension was reported by other researchers. However, the relationship of the volume fraction between the porosity and the solid material shows high divergence among different materials or frozen solvents, as the nature of materials significantly affects the pores formed in freeze casting. Here, we proposed an artificial neural network (ANN) to evaluate the porosity in freeze casting process. After well training the ANN model on experimental data, a porosity value can be predicted from four inputs which describe the most influential process conditions. The error range, reliability and optimization of the model were also analyzed and discussed in this study. The results proved that this method effectively summarizes a general rule for diverse materials in one model, which is difficult to be realized by linear fitting. Finally, a user-friendly mini program based on a well-trained ANN model is also provided to predict the porosity level for customized freeze-casting experiments. Introduction Freeze casting is a versatile and effective colloidal processing technique to produce porous materials, which has experienced fast-growth ever since the 2000s. 1 The technique starts with preparing a stable colloidal suspension, freezing the solvent in the suspension at a controlled low temperature or cooling rate, sublimating the frozen solvent from solid to gas state under reduced pressure, and finally sintering and densifying the remaining solid materials 2 into

218516971

porous materials. 2 Given the fact that the pores are mainly controlled by the frozen solvent, almost any types of solid materials, including ceramics, metals and polymers, can be produced into porous materials via freeze casting. To date, water, camphene and tertiary-butyl alcohol (TBA) are the most reported solvents, which provide a wealth of pore morphologies and structures. Freeze-cast porous materials are promising for a variety of applications including thermal and acoustic insulators, 3 scaffolds, 4-6 advanced composites, 7 gas separation membranes, 8 water filters, 9 electronic sensors 10 and packaging, 11 catalysts, 12 batteries, 13 etc. Porosity is one of the key factors determining the functional and structural properties, such as sound absorption coefficient, 14 thermal conductivity, 15 mechanical strength, 16 etc., of a porous material on top of the solid material characteristics. Freeze casting has been acclaimed as an effective way to control the porosity of the porous materials since the pores generated are, in principle, the replica of the frozen solvent. 2 The linear relationship between the porosity and initial solid materials volume fractions of freeze-cast porous materials have been reported by several researchers based on their process set-ups, specifically. [17][18][19] It leads to the possibility of predicting the porosity of a freeze-cast porous material prior to the freeze casting process, hence saving a considerable amount of time and resource for repeating experiments and processes. However, freeze casting is also considered as a complex process, where various parameters can affect the porosity: formulation of the suspensions (nature of the solid materials, solvents, particle

218516971

sizes, binders, dispersants, etc.), freezing temperature, cooling rate, sintering temperature and time, etc. 1,2 Therefore, the complex interdependent relationships between porosity and these parameters cannot be concluded in a single, isolated paper, as almost every article has its own set of parameters. 1,2 In order to promote better informed experimental design, Dunand et al. 1 analyzed the relationship between the porosity of freeze-cast porous materials and solid material volume fractions among various materials based on more than 800 research papers. Though a linear relationship was proposed, huge diffractions were presented, leading to the difficulty in being used as a reference for the design of new experiments. 1 3 Artificial intelligence (AI) method has a stronger capacity to describe the complicated relationship between variables than traditional statistical methods and empirical methods. One of its advantages is that AI can process different types of data synchronously without barriers. 20,21 Categorical data, numerical data and even images can be integrated into a dataset for a machine learning project. Besides classification and object recognition, numerical regression is one of the objectives which can be realized by machine learning. 22 In the field of materials science, AI, especially machine learning, has been a powerful tool for digging hidden information or potential rules. [23][24][25][26][27] As mentioned above, the volume fraction of the generated pores in freeze casting cannot be accurately predicted from the experimental conditions by simple linear regression because of the different solid materials and solvents characteristics, but it could be a typical problem the regression neural network is able to

218516971

solve. Here, we established an artificial neural network (ANN) to build a projection from the volume fraction of solid materials in the suspension to the porosity of the final freeze-cast porous materials, involving the effects of materials and solvents characteristics. It is a mixeddata neural network which is fed by categorical data and numerical data. The names of materials and solvents are encoded into categorical data, combined with the numerical variable-the volume fraction of solid in the suspension, as the inputs. The corresponding porosity can be quickly predicted from the mixed-data inputs by the ANN model. Our model is trained and tested by the data from real experiments of freeze casting. The architecture and configurations of the network are presented here in detail. In addition, the reasonability and stability of the model was also investigated in this study. Predicting porosity by ANN Artificial neural network (ANN) is a machine learning method which can realize supervised learning. 28 Supervised learning needs a prepared training set with input parameters and associated outputs. In this study, four sets of inputs are selected, which are the type of solid materials, the name of solid materials (e.g. Al2O3), the name of solvents, and the volume fraction of solids in the suspension. We chose three types of materials which are ceramics, metals and polymers, and three types of solvents which are water, tert-butyl alcohol (TBA) and camphene as inputs. The associated output is the porosity of the freeze-cast porous materials. A fully connected neural network was designed for seeking the relationship between

218516971

the inputs and the final porosity. The schematic of the proposed ANN model is shown in Figure 1. The prediction capacity of the ANN model should be validated on the test set and all data (training set and test set) in order to compare with Reference 1 ( Table 1). If some small classifications are contained in the training set, it will cause the imbalance, negatively affecting ANN's learning capacity. 29 Hence, the solid materials with less than 5 samples and the solvents with less than 25 samples have been removed from the raw dataset. The Pearson correlation coefficient (r) is a measure of the linear correlation between two variables. The range of r is [-1, 1], where -1 and 1 mean perfectly negative and positive linear correlation, respectively, and 0 is no linear correlation. 30 The coefficient of determination (R 2 ) is a measure describing how well the model predicts the dependent variable from the independent variable, ranging from [0, 1]. When R 2 is equal to one, it means 100% variation of the dependent variables can be 6 described by the model. 31 For a certain material, the initial volume fraction of solid and porosity has a negatively linear correlation: 17,18,32 , ps ab where ϕp is the porosity, ϕs is the volume fraction of solid, a is the slope, and b is the intercept. However, the regression fitting line cannot describe all the data well with a very small r and R 2 , as reported by Dunand et al. [1] When

218516971

the regression line is obtained from the samples in the same material classification or using the same solvent, the fitting is improved. Less Table S1 (SI). The norm distribution of the error also confirms the reasonability of our ANN model. Figure S2 due to a large amount of data. 8 The results can also be divided into three groups by different types of solvents ( Figure 3). There are 457 samples in the group of water, 76 samples in the group of camphene, and 97 samples in the group of TBA. The errors distribution in three groups can be seen in Figure 3(c), which obtain p ≪ 0.0001 in K-S test. Other evaluation parameters are shown in Table S1 (SI), which prove that the ANN-model has also well-considered the solvent characteristics in learning. In short, our ANN model shows a balanced and excellent performance for all the samples, no matter of which solid materials or solvents are involved. Optimization of ANN Designing the structure of ANN is flexible, and the optimal model is different for different tasks and data types. An unreasonable ANN model will limit the capacity of learning. Therefore, we compared and discussed the ANN models with different loss functions, widths and depths, and other configurations to determinate a reasonable architecture. In addition, the robustness is also confirmed by repetitive training. Loss function ANN aims at minimizing the error during the error propagation. Therefore, a loss function is necessary in ANN to calculate the error, or known as the loss, which can summarize the

218516971

error of all training data down to a numerical value. 34 Table 2. We set the batch size as 12, the learning rate as 1e-4 with a decay of 1e-7, the neuron arrangement as Figure 1 showing, and the epoch as 1000 for all training in Table 2 Depth and width of ANN model 10 The depth and width of an ANN model should be sufficient to process fine features of the data in order to obtain high accuracy. An intact ANN model has at least three layers which are an input layer, a hidden layer, and an output layer. With ANN becoming wider and deeper, the mapping from inputs to output can be more complicated to describe a more abstract relationship. However, an excessively deep ANN is hard to converge, also with problems like gradient explosion and overfitting. 34,35 In addition, the network will be redundant and slow, provided that the width of it is too large. Therefore, we compared the ANN models with different depths and widths in Table 3. We set the batch size as 12, the learning rate as 1e-4 with a decay of 1e-7, the loss function as MSE, and the epoch as 1000 for all training in Table 3. Note that MAE and MAPE can be minimized by a deep and wide enough ANN model (bolded in Table 3). While more neurons are arranged in more than six layers, the performance of the prediction cannot become better. After the systematic comparison, a six-layer ANN model with 128, 64, 32, 16,

218516971

4, 1 neurons in the first to last layer is chosen to process the data of freeze casting. Robustness Only a robust ANN model can describe an authentic and effective relationship between inputs and outputs. 36 Table 4). We set the batch size as 12, the learning rate as 1e-4 with a decay of 1e-7, the loss function as MSE, the neuron arrangement as Figure 1 showing, and the epoch as 1000 for all training in Table 4. The average r is ~0.8, which is much higher than the coefficient between the volume fraction and the porosity (rVF). The average MAE is ~0.07, which is stable in repetitive experiments. Other configurations In addition to the above experiments, other configurations in our ANN model were also investigated respectively. Learning rate is an important hyperparameter which affects the 12 capacity and the speed of convergence. Attempts to set learning rate from 1e-7 to 1e-3 provide the best choice of 1e-4, which ensures the accuracy and meanwhile keeps a relatively high computing speed. We chose leaky ReLU 37 as the activation function for the first to fifth layer. Leaky ReLU with the scalar multiplier of 0.2 shows a slight advantage on prediction accuracy when comparing with ReLU and hyperbolic tangent (tanh). Adam is the most common optimizer for training ANN, 38 also benefitting our study. Batch size is a hyperparameter that defines the number of samples passed to the network for once updating the internal model parameters. It is essential to set a reasonable size for batches, or the

218516971

learning capacity could be weakened. In this study, after various attempts, 12 samples in a batch make the ANN model learn better. After 1000 epoch, most of the experiments we mentioned above show a good convergence, with stable losses in training set and test set. When the ANN model is designed and configured optimally, the loss decreases fast at first and then slowly tries to find a global minimum. The loss variation in 1000 epochs of the optimal ANN model is shown in Figure S2 (SI). The similar losses of the training set and the test set also indicate that the model does not overfit or underfit. Therefore, all models we mentioned in this study are trained for 1000 epochs. Discussion From the above evaluation, a good fit between the predicted data and the experimental data has been achieved. It shows that using inputs such as the type of solid materials, the name Conclusions In Figure 16 S1 (SI). In total, there were 2497 samples, three quarters of them were randomly chosen to feed the ANN model for training, and one quarter of the total were used for testing the trained model. Statistical analysis. In this study, all statistical methods applied to data have been described in detail in the main text, as well as the significant P values. All the statistical analysis was conducted by Python 3.6 with SciPy package. Supporting Information The mini program can be download from: https://github.com/liuyuenus17/FreezeCasting-AI Supporting Information.

218516971

Technical Efficiency of Sole Maize and Maize/Melon Intercrop in Osun State, Nigeria Intercropping has been found out to have enormous economic advantages for the farmer, although with some possible adjoining disadvantages attributable to competition and lack of adequate management. This study therefore analyzed and compared the technical efficiencies of sole maize and maize/melon intercrop farmers in Osun State with a view of determining the factors that influences it. Purposive sampling was used to select 47 respondents each for sole maize and maize/melon intercrop system. Stochastic frontier production function was used to estimate the technical efficiencies. The study revealed that the mean technical efficiency of maize/melon intercrop farmers was higher than that of the sole maize farmers. Costs of land, labor and implement were the factors affecting the sole maize farmers’ efficiency, while the maize/melon intercrop farmers’ efficiency was significantly determined by costs of land, labor and capital. No significant factor determines inefficiency for sole maize farmers. However, land security and extension visit were the major significant inefficiency factors found out to reduce the inefficiencies of maize/ melon intercrop farmers in the study area. It is recommended that famers in the study area should venture into the intercropping of maize and melon more than sole maize cropping; government should strengthen land security and give adequate motivation to extension agents to visit farmers more in order to increase their technical efficiencies. Introduction The growth of two or more crop species simultaneously in the same field during a growing season is known as intercropping [1]. Numerous literatures have often

244675616

argued in favor of one cropping system over the other [2][3][4][5]. It was reported that farmers intercropped for varied reasons [6], including insurance against crop pests, yield increment, weed control and high monetary returns to the farmers. Intercropping is a common practice in tropical developing countries because of its several advantages [7,8] and maize and cassava intercrop has been described as the most prevalent and most productive enterprise with highest net margins in southern Guinea savanna ecology [9]. The cultivation of maize in combination with other crops is a common practice in the tropics [10]. About 73% of the maize in Nigeria is under intercropping [11,12]. Several studies have been carried out on monoculture maize and melon as influenced by intra-row spacing [13,14], The incorporation of 'egusi' melon into maize/cassava intercrop at the right time has been reported to be more profitable and more environmentally friendly [15]. However, it was also reported [16] that the intercropping increased competition between crops and weeds for nutrients and light. In some cases, the output of one crop may be increased through a decline in the production of the other, a situation referred to as 'dominant' and 'dominated', respectively [17]. In some other cases an increase in output of one crop could help to bring about an increase in output of the other crop; this situation is referred to as 'mutual co-operation' [18]. Therefore, the need arises to determine the technical efficiency of sole maize and maize/melon intercrop farmers in Osun State in order to ascertain how efficient each cropping

244675616

system is and compare to be able to determine which has technical advantage over the other in the study area. The objectives of the study were to determine the technical efficiency of sole maize and maize/melon intercrop farmers and factors affecting technical efficiency of sole maize and maize/melon intercrop farmers. Study area The study was performed in Osun State, South-west, Nigeria. It is located between Longitude 7.0° N and 9.0° N, and latitude 2.8° E and 6.8° E with thirty Local Government Areas (LGAs). Osun State is an inland state with Osogbo as the capital. A total of 94 farmers were purposively samples with the assistance of extension agents from Osun State Agricultural Development Program (OSSADEP) from the two LGAs namely: Atakunmo- Annals of Reviews and Research sa West and Iwo based on the predominance of maize and maize/ melon intercrop in these LGAs relative to the rest. Forty-seven questionnaires were administered on each of Sole-maize farmers and Maize/Melon intercrop farmers. Data collection Primary data were collected from respondents with the aid of structured questionnaire. Information collected includes socio-economic characteristics of the respondents, farm characteristics; quantities and prices of inputs and output in the area during the farming season. Methods of data analysis The technical efficiency of both sole maize and maize/melon intercrop and the inefficiency factors were estimated with stochastic frontier production function. The model is specified as follows: where: V i = random or stochastic disturbance term. This captures random variables which are assumed to be identically, independently and normally distributed with zero mean

244675616

and a constant variance where: The technical efficiency of an individual farm is defined in terms of the observed output (Y i ) to the corresponding frontier output (Y i* ) given available technology, that is: So that, 0 1 TE ≤ ≤ . If TE=1, the farm is said to be technically efficient and its output is on the frontier. Otherwise, that is, if TE<1, the farm is technically inefficient because it could have produced more outputs with the given level of inputs irrespective of input prices. The estimates for all the parameters of the stochastic frontier production function and the inefficiency model were simultaneously obtained using the computer program Frontier 4.1 [19]. Results and Discussion Estimates of the parameters in the stochastic production frontier function Table 1 presents the maximum likelihood estimates of the stochastic production frontier model for both sole maize and maize/ melon intercrop farmers. The results revealed the presence of technical inefficiency among both group of farmers in the study area based on the significance of gamma and the Likelihood Ratio The estimated elasticities of the independent variables revealed that costs of land (-1.62), labor (-20.00) and implement (-3.16) were the significant factors affecting the level of production for sole maize farmers. Cost of land was significant at 10% and reduces technical return as its level increases for sole maize farmers. This is because the managerial ability of most of the farmers becomes inefficient with increasing farm size. The costs of labor and implement, both significant at 1%, also reduce the

244675616

level of output with increasing use. This may not be unconnected with the fact that as more labor and implements becomes available, utilization becomes inefficient as a result of excess supply. (***), (**), and (*) denote significance at 1, 5, and 10% respectively, (∆) indicated dummy variables, SE is standard error. For the maize/melon intercrop farmers, the technical efficiency was significantly influenced by cost of land (-3.32), labor (24.93) and capital (14.01). A unit increase in the cost of land, leads to 3.32 reductions in output. For labor and capital, a percent increase in their costs leads to an increase of 24.93 and 14.01 respectively in the level of outcome. Table 1 presents the factors that statistically influence the inefficiency of sole maize and maize/melon intercrop farmers in the study area. For the sole maize farmers, no significant variable which is not consistent with the findings of some studies [20,21], while only land security (-2.17) and extension visit (-3.31) were the major determinants of inefficiency for the maize/melon intercrop farmers in the study area significant at 5% and 1% level. Technical inefficiency analysis This implies that as land security increases, technical inefficiency decreases. Also, as extension visit increases, the technical inefficiency reduces thereby leading to increasing technical efficiency of the maize/melon intercrop farmers. Technical efficiency analysis From Table 2, the mean predicted technical efficiency for the sole maize farmers was 69% lower than that of the maize/melon intercrop farmers (75%) and that on average, maize/melon intercrop farmers were more technically efficient than the sole maize farmers.

244675616

This implies that sole maize farmers were 31% inefficient more than the inefficient maize/melon intercrop farmers (25%). This is in tandem with recent findings [4,15,22]. However, this finding negates another research [23] that in maize-melon mixtures, the sole crop components yielded higher than the corresponding crops in intercropping situation. For an average sole maize farmer that operating at the highest level of efficiency, such farmer is saving a cost of 31% (i.e. ). However, majority of sole maize farmers (51.06%) had technical efficiencies between 71-100% while the percentage for the maize/melon intercrop farmers was 59.58%. Conclusion and Recommendations The study brought to fore the differences in level technical efficiency of sole maize and maize/melon intercrop farmers in the study area, more importantly as a result of inefficiencies in the farming system employed. The study revealed that the mean technical efficiency for the maize/melon intercrop farmers was higher than that of the sole maize famers, meaning higher efficiency. Costs of land, labor and implement were the factors affecting the level of sole maize farmers' productivity, while the maize/melon intercrop farmers' productivity was significantly determined by costs of land, labor and capital. Also, land security and extension visit were the major significant inefficiency factors found out to reduce the inefficiencies of maize/melon intercrop farmers in the study area. Based on the findings of this study, the following were recommended for the farmers in the study area. 1. It is advised that famers in the study area should venture into the intercropping of maize and melon more than sole

244675616

maize cropping. 2. The rights to land should be strengthened to enable maize/melon intercrop farmers enjoy better technical returns from the land. 3. Also, extension agents should make more visits to the farmers, bringing more innovations and information, to ensure that farmers achieve a higher level of technical efficiency.

244675616

Effect of temperature changes on antioxidant enzymes, and oxidative stress in Nerita oryzarum collected along Atomic Power Plant sites Effect of heated euent released from the Tarapur Atomic Power Station, India were investigated by analyzing the level of antioxidant enzyme activities on Nerita oryzarum. Seasonal variation of antioxidant enzyme (LPO, CAT, SOD, GPx and GST) were determined at six location. CAT increases in the pre-monsoon season (range of 0.12 to 2.1 mM of H 2 O 2 consume/min/mg of protein) and LPO activities also increasing trend (range of 0.4 to 2.3 nM of MDA/min/mg) during same pre-monsoon at NIII and SII respectively. Enzyme values encountered high during pre-monsoon, which indicating that increased in temperature is resulting in increased activities. In experimental condition also (30 0 C, 35 0 C, 40 0 C and 25 0 C as control), antioxidant enzyme activities were in increasing trend due to raises of water temperatures. Present study was prima facie work related to physiological response of N. oryzarum related to heated euent released from atomic power station and useful as baseline information for future research work. Introduction To generate the energy in living body, consumption of oxygen through breathing is needed. Most of the oxygen gets incorporated into harmless water molecule except up to 2 percent may be spun off in to our cell as free radical or ROS (Reactive Oxygen Species), which are greatly charged molecules. The production of ROS are dominant in aquatic ecosystem. The oxidative stress are very signi cant component as stress response, especially in marine

236406758

environment were intertidal organisms get exposed to assortment of environmental as well as anthropogenic stressors, for example thermal related stress which is very predominant due industrialization, power plants and global climate change. Oxidative stress can brie y de ne as, the production of ROS beyond the ability of organism to reduce and be able to cause damage to proteins, lipid and DNA (Halliwell and Gutteridge, 1999). The quenching of 1 O 2 on site of production and to decrease the uctuation of other ROS to ultimately inhibit the production of furthermost damaging ROS, i.e., HO. is the primary function of the antioxidant defense system (Cadenas, 1989;Fridovich, 1998). Organisms in marine ecosystem usually get exposed to multiple environmental stressors, such as increase in temperature, reduction in pH, uctuation in salinity and exposure to metals (Harley et al., 2006;Brierley and Kingsford, 2009;Sparks et al., 2019;Albayrak et al., 2019). Especially the organisms from rocky intertidal and shallow water areas have more negative impact (Millero, 2007) than pelagic zone organisms (Beuchel et al., 2006). Increase in temperature result in elevated activities of cytosolic and mitochondrial enzyme system that generates ROS in extra mitochondrial compartment which contribute to both increase rate of metabolism and ROS production (Sokolova et al., 2012). High temperature increases mitochondrial ROS production in mollusks (Heise et al., 2003;Abele et al., 2002). A moderate increase in temperature leads to stimulation of the antioxidant protection system and expression of antioxidant enzymes get induced (Ivanina et al., 2009;Gourgou et al., 2010). Since, gastropods are slow moving and inhabitant of intertidal

236406758

regions, suggesting that they already get exposed to thermal stress due to exposure during tidal rhythm, any additional thermal stress, such as, heated e uent from nuclear power plant, may cross their optimum tolerance limit. In the present study, the gastropods of Nerita oryzarum was selected the study animal due to availability of all seasons around Tarapur Atomic Power Station, India and it could be the greatest demonstrative model for studies of in uence of stress on antioxidant defense mechanism. Results 2.1 Measurement of lipid peroxidation level due to temperature changes The oxidative lipid degradation is known as lipid peroxidation. It is the development in which free radicals take electrons from the lipids in cell membranes, nally cells got damage. These process done by mechanism of free radical chain reaction. During present study the lipid peroxidation in whole body tissue of N. oryzarom was analysed spatially at six different locations as well as seasonally during six different seasons in close vicinity of nuclear power plant site, Tarapur (S Table-2). During the monsoon the minimum LPO activity was measured of 0.58 nM of MDA/min/mg at S I and maximum of 2.31nM of MDA/min/mg at SII location with (STDV) ± 0.41. Post-monsoon season the LPO activity measured in N. oryzarum got reduced in the range of 0.09 (at NIII) to 1.21 (at SII) nM of MDA/min/mg with (STDV) ± 0.29. During the pre-monsoon it again showed increasing trend and was in the range of 0.4 (at NIII) to 2.3 (at SII) nM of MDA/min/mg with (STDV) ± 0.5

236406758

(Fig. 1). Catalase analysis Catalase is present in all aerobic organisms. Within the cell, catalase in mainly located in peroxisomes and mitochondria as both soluble and membrane bound form. Catalase in alternative dismutase enzyme comprises heme moiety at its active position. It changes H 2 O 2 in H 2 O and molecular oxygen mostly in peroxisomes. It avoids conversion to hydroxyl radical and other toxic ROS. In the present study the CAT was measured in the range of 0.03 (at NI and NII) to 1.1 (at SI) mM of H 2 O 2 consume/min/mg of protein with (STDV) ± 0.35 at monsoon season. During the post-monsoon season CAT values does not showed any trend and was in almost similar range as of monsoon season i.e., 0.03 (at SI) to 1.1 (at SII) mM of H 2 O 2 consume/min/mg of protein with (STDV) ± 0.34 (S Table-2). The CAT increases in the pre-monsoon season and was in the range of 0.12 (at NIII) to 2.1 (at SII) mM of H 2 O 2 consume/min/mg of protein with (STDV) ± 0.42 (Fig. 2). Effect of temperature changes on SOD level Superoxide dismutase catalyzes the dismutation of superoxide (O 2 − ) radical into hydrogen peroxide (H 2 O 2 ) or ordinary molecular oxygen (O 2 ) providing cellular defense against reactive oxygen species. It was observed that the minimum activity of SOD was recorded at NII location during all the season with minimum of 1.8 U/mg of protein during post-monsoon followed by 2.3 U/mg of

236406758

protein in monsoon and 3.5 U/mg of protein in pre-monsoon season. The maximum values of SOD activity was recorded at SII location throughout study period with similar trend i.e., minimum of 13.4 U/mg of protein in postmonsoon followed by 19.2 U/mg of protein in monsoon and maximum of 27.5 U/mg of protein during pre-monsoon season (Fig. 3). Glutathione peroxidase is selenium comprising antioxidant enzyme that e ciently reduces H 2 O 2 and lipid peroxides to water and lipid alcohol, one-to-one and in turn oxidizes glutathione to glutathione disul de. The Glutathione peroxidase helps to prevent lipid peroxidation and maintains intracellular homeostasis as well as redox balance. The GPx values were minimum during the post-monsoon season and it was in the range of 0.04 (at NII) to 1.1 (at SII) nmol glutathione oxidase /min/mg protein with (STDV) ± 0.24. During monsoon it was in the range of 0.4 (at NIII) to 2.154 (at SI) nmol glutathione oxidase /min/mg protein with (STDV) ± 0.37 and in pre-monsoon it varied in the range of 0.4 (at NIII and SIII) to 3.04 (at SII) nmol glutathione oxidase /min/mg protein with (STDV) ± 0.58 (Fig. 4). Glutathione S-transferases level Glutathione S-transferases (GST) are a group of enzymes that catalyze conjugation of reduced glutathione to a varied range of substrates, nally resulting in detoxi cation (Graham et al., 2013). They also function as transport proteins. It catalyses the conjugation of GSH via a sulfhydril group to electrophilic centers on wide variety of substrates to make the compound more water soluble. The

236406758

GST values showed the increasing trend; minimum during post-monsoon followed by monsoon and premonsoon. It was minimum of 8.9 (at SIII) nmoles/min/mg Protein during post-monsoon season. During the monsoon it varied in the range of 10.2 (at NIII and SIII) to 17 (at NI) nmoles/min/mg Protein and it increased in pre-monsoon with the range of 10 (at SIII) to 20.1 (at SII) nmoles/min/mg Protein with (STDV) ± 2.09 (Fig. 5). Similarly, all the antioxidant enzyme activity was studied in experimental organisms and the result showed that, all the activity was in increasing trend and directly proportional to the water temperature. The result of experimental analysis is given S In the present study the in uence of heated e uent water released from the Atomic plant at Tarapur were investigated by examining the antioxidant enzymes as molecular biomarkers. The ve different assay's were analysed such as Lipid peroxidase, Catalase, Superoxide dismutase, Glutathione-S-transferase and Glutathione peroxidase in N. oryzarum collected from six different locations during three different seasons. The thermal impact was also studied in similar organism at laboratory condition for similar biomarkers as at environmental conditions by maintaining the organism at three different temperatures (30 0 C, 35 0 C, 40 0 C and 25 0 C as control). The correlation of these biomarkers with response to heated e uent was studied (Table-2). Discussion All aquatic animals, their diversity and community structure are in uence by temperature. Various studies have suggested that the temperature bring out considerably the physiological stress and higher metabolic activity which increases oxygen

236406758

consumption and consequently produces ROS (O 2 − , H 2 O 2 and OH − ) (Lushchak and Bagnyukova, 2006;An and Choi, 2010). The temperature limits are different for different species within which they can successfully survive, for some species the range of this limit is very narrow. The metabolic scope and stress response mechanisms are primarily important factors that allow the survival of particular species in that range limit (Somero, 2002 andPortner, 2010). To maintain the physiological homeostasis within the temperature range various cellular mechanisms plays a vital role. Off this the important are, heat shock proteins, mechanism of general alteration of energetic metabolism and most primitive, antioxidative system (Arad et al., 2010, Pohlmann et al., 2011,Qian et al., 2012and Mizrahi et al., 2012. According to Handy et al. (2009), marine invertebrate's response to temperature by adjusting mitochondrial electron transport, which in turn produces the superoxide anions as a bi-product (Davidson and Schiestl, 2001). The antioxidant enzyme activity uctuates as per the production and scavenging of ROS produced. The organism is said to be in oxidative stress when the antioxidant defense mechanisms get oppressed by production of reactive oxygen species (Menghong Hu et al., 2015). The use of biomarkers in environmental monitoring programs, both in freshwater and marine water, frequently uses molluscs as sentinel species (Marigómez et al., 2006;Laffon et al., 2006;Cajaraville et al., 2006;Lima et al., 2007). The antioxidant defense mechanism of marine mollusks has been the subject of many researchers. Many studies mostly assessed the response of antioxidant defense system of mollusks

236406758

related to speci c pollutants, such as metals (Nikitha Divakaran, 2014), polycyclic aromatic hydrocarbons (Sarkar et al., 2008;Bhagat et al., 2016), oil spill (Moreira et al., 2004;Ana et al., 2009), organic-inorganic pollutants, Industrial and domestic waste disposal etc. In particular most of study focused on bivalve species (for review, Winston and Di Giulio, 1991). But little is known about antioxidant defense of marine gastropods, especially the data related to thermal stress due to heated e uent released from nuclear power plant sites is seldom. Present study investigates the physiological response of N. oryzarum in the form of response of antioxidant system towards heated e uent released from nuclear power plant site, Tarapur. Antioxidant enzymes of SOD, CAT, GPx, GST and GR are signi cant in coping with oxidative stress produced by metabolism itself or/ and environmental factors also (Tripathi et al., 2006;Pinto et al., 2003). The antioxidant system in mollusks involves some important enzymes which prevent oxidative damage mediated by ROS are Catalase, Superoxide dismutase, Glutathione-S-transferase and Glutathione peroxidase. In the present study most of the enzymatic activity and lipid peroxidation values encountered high during pre-monsoon and at SWIII location, which indicating that increased in temperature is resulting in increased activity. These results are in correlation with experimental condition where increased in temperature also leads to increased in the activity of antioxidant enzymes. Many authors studied the effect of temperature, pH and salinity on antioxidant systems of mollusks. Matozzo et al., 2013, studied the combined effect pH and temperature on bivalve and suggested that exposure of

236406758

mollusks to the ROS generating stress full condition may either induce or inhibit the activity of antioxidant enzymes. Lipid peroxidation can generally be described as the process in which the carbon-carbon double bond in lipid especially PUFA (Polyunsaturated fatty acids) get attacked by oxidants such as free radicals (Antonio et al., 2014). In the present study TBARs test was used to represent the level of LPO in Neritta oryzarum. The maximum LPO activity (1.88 ± 0.12 nM of MDA/min/mg of protein) was estimated during pre-monsoon 2014, when the maximum water temperature was recorded, indicates that, increased in temperature induces the LPO activity. Production of free radicals increased with increase temperature and it reacts with polyunsaturated fatty acids of the cell which induces the LPO activity (Jacky Bhagat et al., 2016). Under experimental treatments, the LPO activity was directly proportional to temperature (Table-1) and con rms the present nding. The similar result was also reported by Verlecar et al., 2007, in Perna viridis exposed to high temperature. Superoxide dismutase accumulates in response to oxidative stress and is a main antioxidant defense pathways (Down et al., 2001). Increased SOD level shows that cells are answering to an oxidative stress (Fridovich, 1995). SOD being the primary scavenger of O 2 . − during heat stress indicates that it inhibiting accumulation of oxygen radicals. Verlecar et al., 2007 reported increased activity of SOD in digestive gland and gills of Perna viridis at higher temperature (32 0 C) than the control condition (26 0 C). Increased in temperature substantially increase the

236406758

oxygen consumption in Nacella concinna (Peck, 1989) and also leads to increased SOD activity in digestive gland and gills (Abele et al., 1998). de Oliveira et al., 2015 have also reported the similar trend in N. concinna gastropods that increased in temperature leads to increased SOD activity. In the present study maximum SOD activity was detected at SW II location during pre-monsoon 2014 (22.57 ± 0.89 U/mg protein) when the temperature was recorded highest and was in the range of 28.2 0 C to 38.5 0 C (Avg. 31.20 0 C ± 3.78) followed by monsoon 2013 (16.71 ± 0.57 U/mg protein), where temperature was in the range of 28 0 C to 33.75 0 C (Avg. 30.71 0 C ± 1.99) indicating that increased in temperature leads to increased in SOD activity (although not linear in organism from natural habitat). The result was con rmed in experimental set up where organism was exposed to increased temperature for 8 hrs leads to an increased SOD activity (Table-1). This results is similar with Matozzo et al., 2013, encountered the upward trend of SOD activity with increased in temperature in clams and mussels. Parihar et al. (1997), have also described that, increased in SOD activity in gills of H. fossilis on short period temperature improvement. In present study the SOD values shows the seasonal as well as spatial trend. It was minimum during post-monsoon 2013 and maximum in pre-monsoon 2014. Spatially, intake locations i.e. NWI and SWI had less value for SOD activity than the discharge locations. SOD

236406758

values shows positive correlation with pH (r = 0.439) and water temperature (r = 0.911) and also it is positively correlated with CAT (r = 0.798) and LPO (r = 0.784). M. Hu et al., 2015 also have similar results and stated that SOD activity was high in gills and digestive glands of M. coruscus kept at low pH and high temperature. Increased in SOD activity carry out the procedure of O 2− dismutation to H 2 O 2 which should ultimately stimulate the activity of Catalase enzyme, which then bring out the reduction of H 2 O 2 and protects the tissue from ROS and hydroxyl radicals. Thus catalase act as very important scavenger in the cell. In present study the activity of catalase was recorded high at SWII location (1.52 ± 0.14 mM of H 2 O 2 consume/min/mg of protein) during pre-monsoon 2014 (Figure-3). Under the oxidative stress often activity of catalase also increases (Hermes- Lima, 2004), this is in consistent of current nding indicating increased in temperature also indices the activity of SOD and CAT. The result was also similar during the experimental condition where organisms were exposed to thermal stress also have shown similar trend of increased in catalase activity from 0.41 ± 0.2 mM of H 2 O 2 consume/min/mg of protein to 1.17 ± 0.3 mM of H 2 O 2 consume/min/mg of protein at 30 0 C and 40 0 C respectively (Table-1). The catalase activity was higher in experimental thermal set up (i.e. 30 0 C< 35

236406758

0 C< 40 0 C) than in control (25 0 C, CAT = 0.39 ± 0.1 mM of H 2 O 2 consume/min/mg of protein) con rmed that organism exposed to thermal stress induces the higher activity of catalase in order to cope up with oxidative stress. Abele et al., 1998, have demonstrated the thermal stress in Nacella concinna by exposing it to increased temperature for 24 hrs and concluded the increased in SOD and CAT activity in the gills. In the present nding CAT activity showed positive correlation with SOD activity (r = 0.798) and with water temperature (r = 0.911) and pH (r = 0.439) indicating both parameters are in uencing the enzymatic reactions in N. oryzarum (Table-2). Along with catalase, glutathione peroxidase also acts as important scavenger in the cell. GPx are a large family of selenium dependent and selenium independent enzymes with peroxidase activities. Reduction of organic hydroperoxides to their corresponding alcohol and free H 2 O 2 to water are the main biological roles of GPx. In cellular matrix it is located in cytosol, mitochondrial matrix and membrane, thus it removes H 2 O 2 from various cellular compartments in combining with catalase (Halliwell and Gutteridge, 2007). During present study, the highest activity of GPx was measured at SWII location (1.98 ± 0.2 nmol glutathione oxidase/min/mg protein) during pre-monsoon 2014, followed by monsoon 2013 at NWII location (1.39 ± 0.1 nmol glutathione oxidase/min/mg protein). Both locations are the discharge point of TAPS 1&2 And TAPS 3&4, respectively and seasons when maximum

236406758

values for water temperature was reported during study period, indicates that the increased in water temperature in uence the activity of GPx. Similarly, when the organism was exposed to thermal stress also shows directly proportional relationship between GPx and water temperature. In the present study GPx activity was positively correlated with water temperature (p = 0.808), SOD activity (p = 0.704) and salinity (p = 0.449) ( Table-2). Glutathione S-transferases (GST), enzymes catalyses the conjugation process of condensed glutathione to extensive range of substrates and ultimately results in detoxi cation. It catalyses the reaction of glutathione compound with -SH group, thereby neutralizing their electrophilic sites and making the compound more water soluble. Temperature signi cantly affects the GST activity (Lynn et al., 2000), and it increases with increase in temperature. In present study the GST activity was maximum (16 ± 0.57 nmoles/min/mg protein) was measured during pre-monsoon 2014. At experimental condition the GST activity increased with increase in temperature from 30 0 C to 35 0 C and it reduces at 40 0 C, these could be reason that the organism has already cross the lethal temperature and further response of antioxidant system has stopped. Although, GST is frequently used as biomarker in many experimental studies, the systematic studies of its kinetics, ecologically relevant endpoints and affect of environmental factors are limited. In case of invertebrates, the Glutathione S-transferase activity does not give any predictable trends, because detailed study of its kinetic behavior under stress condition and characterization of several enzymes with in this family is

236406758

seldom (Domingues et al., 2010). Biomarkers, when used as diagnostic tool for environmental monitoring, it is important to link the response of biomarker at individual level, potential effect of pollutant on population, ecosystem and on communities of particular organism. However, often it is di cult as meaningful interpretation required deep knowledge of integration of various levels of biological organization and its mechanistic processes (Forbes et al., 2006). Moreover, one should also consider that spatial and temporal history related to particular contaminants which in uence the biochemical changes in organisms, and if these changes are in physiological response and ecologically relevant, then biomarkers are successfully acts as measure of effect (Ankley et al., 2007). Sample collection The Nerita oryzarum (around 20-25 organisms from each site) were collected during three different seasons (monsoon 2013 to pre-monsoon 2014) from six rocky intertidal locations (TAPS1&2, Light house, Varor, TAPS 3&4, Popharan and Nandgaon) around Tarapur nuclear power plant site (S Table-1 & S Fig. 1). The gastropod of similar size (23 to 32mm) and irrespective of their sex was collected for analysis. The organisms were freeze in liquid nitrogen at the site of collection and transported to laboratory. In laboratory, shell of gastropod was gently smashed and entire body tissues were cautiously dissected from other vicera. A soft tissue from 5 to 6 gastropods were pulled together to make 1gm of weight, which was used for preparation of tissue extract for biochemical assay. Analaysis of protein and lipid peroxide (LPO): The estimation of protein in N. oryzarum, The Lowry's method

236406758

(Lowry et al., 1955). The standard procedure for estimation of level of lipid peroxide in organism tissue by their reaction with thiobarbituric acid developed by Ohkawa et al., 1979 was used. To estimate lipid peroxidation, 1 g of tissues was homogenized in 9 ml of ice cold 0.25 M sucrose. 0.2 ml of animal tissue homogenate was assorted with 0.2 ml of 8 % SDS followed by 2 ml of 20 % acetic acid and 2 ml of 0.8 % Thiobarbituric Acid (TBA). Further, the mixture was make up to 4 ml by using distilled water and heated for 60 min in boiling water bath. After cooling the mixture was centrifuge at 3000 rpm for 10 min. The absorbance was measured at 532 nm and the results were expressed as nM of MDA min − 1 mg − 1 of protein. Lipid peroxide was considered as MDA equivalent. Determination of Catalase (CAT): The catalase assay was measured according to simple calorimetric method for rapid assay of catalase described by Sinha (1972). Tissue homogenate for catalase activity was prepared in 4 ml of phosphate buffer (0.1 M, pH 7.4) by using tissue homogenizer. The total contents were centrifuged at 4 0 C, 18,000 rpm for 30 min. The reaction mixture contents 0.5 ml of 0.2 M H 2 O 2 (Hydrogen Peroxide), one ml of 0.01 M sodium phosphate buffer (pH 7.0) and 0.4 ml of double distilled water. Further, two ml of dichromateacetic acid reagent, which contains glacial acetic acid and potassium dichromate and in 3:1

236406758

proportions, were added to stop the reaction. The reaction mixture heated for 10 min. in water bath and further kept for cool. After cooling, green color was developed and the absorbance was measured at 583 nm and blank also maintained. The CAT activity was expressed as mM of H 2 O 2 consumed/min/mg of protein. Determination of Superoxide dismutase (SOD) The SOD (Superoxide dismutase) activity was measured by the method of Misra and Fridovich (1972) with minor modi cation. The SOD activity in tissue homogenates were analysed by the level of autooxidation of epinephrine to adrenochrome. For SOD determination, 1 g of tissue was homogenized in 0.1M phosphate buffer (pH 7.4), 0.2 ml of chloroform (ice chilled), 0.15 ml of ethanol (ice chilled) followed by 1 ml of distilled water was added in chilled condition. The total contents were centrifuged for 10 min at 4 0 C and 3000 rpm in cooling centrifuge machine. The one ml of reaction mixture contained epinephrine (10 mM), sodium carbonate buffer (50 mM with pH 10.2) and EDTA (10 mM). The absorbance was measured for 3 min at 30 sec interval at 320 nm. The SOD activity results were reported as U mg − 1 (per mg of protein), where 1 U of SOD is distinct as the quantity of sample causing 50 % of inhibition of epinephrine auto-oxidation. Determination of glutathione peroxidase (GPx) The estimation of glutathione peroxidase in tissue homogenate was done by following the method of Paglia and Valentine (1967) The activity of GSH (Glutathione-S-transferase) measured by

236406758

the method describe in Habig et al., (1974). The reaction mixture content 0.1 ml of tissue homogenate mixed with 1 ml of phosphate buffer (0.1 M K 2 HPO 4, EDTA-Na 2, pH 6.5) followed by 0.1 ml of CDNB (1-Chloro-2,4-Dinitrobenzene). Further, make up the volume of reaction mixture to 2.5 ml by using double distilled water and incubated at 37 0 C for 5 min. Then add 0.1 ml of 30 mM reduced glutathione (GSH). The change in absorbance was measured at 340 nm for 3 min at 15 sec interval. The reaction was based on the conjugation of CDNB solution and reduced glutathione reaction in reaction buffer. The GST activity was measured by using the extinction coe cient of 9.6 mM − 1 cm − 1 for CDNB. The results were stated as µmol/min/mg of protein. 4.7 Experimental set-up for analyzing antioxidant enzymes changes, and oxidative stress in Nerita oryzarum due temperature variation The gastropod species Nerita oryzarum were collected at rocky intertidal locations from Varor which was considered as reference location, as it is situated away from impact of heated e uent. The gastropods of similar size (18 to 20 mm) and irrespective of their sex was collected for our experiment study. The N. oryzarum was brought to library in a plastic bucket half lled with sand and seawater from their natural habitat. While collecting the animal, the temperature of sea water was measured (29.7 ± 1 0 C). In laboratory the organisms were transferred carefully to the glass tank of 20 L

236406758

capacity (S Fig. 2). The sand from natural habitat of organism was spread at the bottom of a tank. The organisms were allowed to acclimatize in aerated seawater for 24 hours. Then the waterproof heater was inserted to each tank after setting it to desire temperature ranges. The 30 0 C was directly set in heater and for two other temperature ranges it was slowly increased with an interval of 2-3 0 C until the temperature of seawater in tank reach 35 0 C and 40 0 C. The water temperature, pH, dissolved oxygen and salinity were recorded from 0 hr. time to 8 hr at every 2 hr interval. The acclimatized N. oryzarum samples were divided in three groups and each group had three replicates as tank A, B and C. In each tank, 8-10 animals were placed. Have three tanks labeled as A, B, C (maintained at 25 0 C temperature) details are mentioned in Table-1. The animals were sacri ced after 8 hrs of experimental duration. The temperature range and duration was xed based on led observations. Since, the N. oryzarum collected are inhabitant of intertidal region which are exposed to tidal rhythm of around 6 to 7 hrs of diurnal variation. Statistical analysis: The data received from all analytical methods were treated statistically using ORIGIN software. Two-way ANOVA with Duncun's test were run by using SPSS v.20. The graphs were plotted by using SIGMA PLOT software. The correlation analysis between environmental parameters and antioxidant values were analysed on STATISTIKA 7 software. Expressive

236406758

data analysis were performed, including the calculation of mean, SD, maximum and minimum. Conclusion The results of present study is prima facie work related to physiological response of gastropod of N. oryzarum related to heated e uent released from atomic power station and useful as a baseline information for future research work. Further this study provide a meaningful understanding of how temperature changes affect the antioxidant defense system in marine intertidal organisms. Studies on mechanisms for individual function of enzymes in temperature induced stress nding also required in future. Declarations

236406758

NEW HEURISTIC ALGORITHMS FOR ROLLING AIR FRAME AERODYNAMIC PARAMETERS ESTIMATION In this research, based on heuristic optimization algorithms, three new strategies are developed for Aerodynamic Parameters Estimation (APE) of one pair ON-OFF actuator rolling airframe. In the 1st method namely EAM-PSO the aerodynamic parameters are directly estimated. While, the next two algorithms called EBM-PSO and SEBM-PSO are twostep strategies. In the 1st step the aerodynamic forces and moments are estimated, then after passing through a designed smoothing filter, in the 2nd step aerodynamic parameters are estimated. In EBM-PSO all the aerodynamic parameters are estimated at once by solving one optimization problem. In SEBM-PSO the APE is converted to solve four separate optimization problems. A modified particle swarm optimization algorithm is developed and used in estimation process. The performance of proposed algorithms is compared with that of state of the art algorithm EKF. The simulation results show that SEBM-PSO and EBM-PSO are better than EAM-PSO in term of accuracy and run time. Introduction An accurate aerodynamic model is crucial for developing high fidelity flight simulations and designing efficient control systems. Computer-based methods like theoretical-empirical relations and computational fluid dynamic (CFD) method as well as experimental approaches including wind tunnel and flight test are common ways to estimate the aerodynamic characteristics. The theoreticalempirical methods have a low accuracy relating to some simplifications done for their relations derivation. CFD methods are time consuming. Both of computer-based methods are sufficient for first design steps, but for more accurate results, we need to use experimental approaches. The wind tunnel test approach,

218815881

itself, has uncertainties arising from scaled models and wind tunnel walls effect. It also has limitations in simulating the vehicle's whole flight envelope. Flight test method is the most accurate approach for all flight conditions. Therefore, during last decades, several approaches have been developed for APE based on flight test data. These approaches can be categorized as classic filter-based approaches, intelligent algorithms, and heuristics optimization methods. There are various researches in literature utilizing classic filters for APE problem. The observer/Kalman filter identification method is applied to the problem of online system identification of an aircraft (Valasek & Chen, 2003). Based on real flight test data, three recursive parameter estimation algorithms including EKF, simplified Unscented Kalman Filter (UKF) and UKF are compared for APE of two aircrafts (Chowdhary & Jategaonkar, 2009). EKF based technique for the calibration of various parameters measured by air data system of an aircraft is studied in (Kamali et al., 2016). In Bayoğlu and Nalci (2016) an adaptive KF is applied to APE for a supersonic missile having rapid speed variations. A two-step methodology, studied in Moszczynski, Leung, and Grant (2019), in which an adaptive maximum aposteriori trajectory estimation scheme is adopted for accurate dynamic model identification. Two concepts namely Estimation After Modeling (EAM) and Estimation Before Modeling (EBM) are introduced for the first time by (Mohammadi & Massoumnia, 2000). According to EAM strategy, the aerodynamic parameters are directly estimated from the measured data. This is the most common approach have been used by researchers. While, in EBM method two extra sensors are

218815881

needed and estimation is done in two steps. In that article, in first step, aerodynamic forces and moments were estimated using Frazier smoother, and in second step, APE is done using EKF. Classic filters are more convenient but they have some limitations: Kalman filter is optimal only for linear systems with Gaussian noise. EKF and UKF methods are used for nonlinear systems, but the problem associated with them is greatly attributed to their needed approximations in linearization process (Moszczynski, Leung, & Grant, 2019). Intelligent algorithms like fuzzy and neural networks methods are knowledge-based techniques, can do tasks based on the data given for training or previous experiences. The learning process itself may take a very long time and there is usually no guarantee to success (Kumari, Sunita, & Smita, 2013). Intelligent algorithms like neural network and fuzzy control are also applied to the APE problem. Modeling and identification of spacecraft system using adaptive neuro-fuzzy inference systems is performed in (Hanafy, Al-Harthi, & Merabtine, 2014). Hatamleh et al. (2015) have focused on estimating unknown dynamics model parameters of an unmanned quadrotor under presence of noisy feedback signals, where iterative bi-section shooting method, artificial neural network method and a hybrid approach are used. In current decade, nature inspired heuristic optimization algorithms methods are also implied to the parameter estimation problem. A review of intelligent algorithms including genetic algorithm, particle swarm optimization (PSO) and artificial bee colony algorithms is proposed in Wang et al. (2013) for aircraft APE. Taboo continuous ant colony system method (Nobahari & Sharifi, 2014), is

218815881

used in Rezaei (2015) to estimate aerodynamic coefficients of a rolling airframe. Propeller z-force and pitching moment coefficients of a small unmanned aerial vehicle are estimated through a modified PSO in Tieying, Jie, and Kewei (2015), which in, accelerations, deflection angle, and pitch rate are taken as observations. A new optimization algorithm called adaptive chaotic mutation PSO is proposed to perform APE for a spinning symmetrical projectile (Guan et al., 2016), where, only aerodynamic drag and lift coefficients are estimated. Bian et al. (2016) presented an improved PSO method, which in every particle will do a local search before updating the global velocity and position. Then, the global best particle is created by a certain number of elitist particles. This algorithm is used to identify stability and control derivatives for an experimental small unmanned helicopter. The advantage of heuristics algorithms is their global search capability. But, they suffer from high computational time problem (Sone & Yadav, 2015). The current research concentrates on developing new APE algorithms for a one pair ON-OFF actuator rolling airframe, based on heuristic optimization approaches. PSO algorithm is selected because of its capabilities in finding global optima in continuous optimization problems. Most APE algorithms are based on the conventional strategy namely EAM. In the current research, a new EAM based PSO algorithm, namely Estimation After Modeling Particle Swarm Optimization (EAM-PSO) is developed. In addition, two novel strategy namely Estimation Before Modeling Particle Swarm Optimization (EBM-PSO) and Separated Estimation Before Modeling Particle Swarm Optimization (SEBM-PSO) are developed and applied to the APE problem.

218815881

EBM-PSO algorithm is a two-step strategy: In 1 st step, the aerodynamic forces and moments are estimated, then after passing through a designed smoothing filter, in 2 nd step, aerodynamic parameters are estimated. In EBM-PSO, the aerodynamic parameters are estimated via solving a single optimization algorithm based on a modified version of PSO. The derived formulation makes it possible to eliminate the numerical integration process and hence decreasing the computational time. Utilizing the symmetrical properties of rolling airframes, the aerodynamic parameters can be estimated by solving separate optimization algorithms. This resulted in SEBM-PSO algorithm. SEBM-PSO provides more exact results. The performance of newly developed algorithms and EKF, as a state of the art algorithm, are compared. It is shown that SEBM-PSO and EBM-PSO algorithms provide more exact results in comparison to those of EAM-PSO and EKF algorithms. EBM-PSO and SEB-PSO also have better run time than EAM-PSO because of eliminating the numerical integration during optimization algorithm iterations. This paper is organized as follows: In section 1, the rolling airframe equations of motion are illustrated. Section 2 is devoted to the proposed modified particle swarm optimization (PSO) algorithm. In section 3, EAM-PSO method in addition to EBM-PSO method are explained in details. Simulation results are discussed in section 4, and finally conclusions are presented. Airframe equations of motion The vehicle of concern is a roll-stabilized airframe with one pair ON-OFF actuator. For modeling this vehicle, it is assumed that the thrust vector is fixed through the center of gravity and coincide with the body frame x-axis, the

218815881

airframe is rigid, Vehicle has an axis-symmetric and cruciform shape, so that the moments of inertia I yy (t) and I zz (t) are identical and products of inertia moments can be discarded. Since, vehicle is modeled as symmetric in pitch and yaw planes (Aksu, 2013), the aerodynamic coefficients C y and C n are identical to C z and C m respectively. Therefore, we have; , , , Taken into consideration the above assumptions, the equations of motion in the body-fixed frame are as follows; cos sin sin tan cos tan d p q sin cos cos cos where, u, v, and ware velocity components in body-fixed coordinate system, p, q, and r are vehicle angular rates, ϕ, θ, and ψ are Euler angles, D is body diameter, S is a reference area and Q is the dynamic pressure. In presence of wind [u wind , v wind . w wind ], total velocity V T , angle of attack α, and sideslip angle β can be determined as follows (Vitale, 2013): Here, it is assumed that wind effect is negligible in comparison to vehicle velocity and one can say: Modified particle swarm optimization algorithm The original PSO, first developed by (Kennedy & Eberhart, 1995), is inspired by the social behaviors observed in flocks of birds and schools of fish. The main feature of PSO is that it works in continuous search spaces. Considering the algorithm proposed by (Karimi, Pourtakdoust, & Nobahari, 2011), which is a PSO-based constrained optimization algorithm, a non-constrained version of this

218815881

algorithm is developed here. It starts with an initial population of randomly generated particles. For a search problem in a D-dimensional space, a particle represents a potential solution presented by its velocity and position. During a search process, each particle is attracted by its previous best particle (X best ) as well as its neighborhood best particle (X lbest ). The swarm uses two equations to update each particle's velocity and position. ( ) ( ) where, V i,j and X i,j are the velocity and position of particle i in j th dimension of search space, k = 1,…, N itr is number of iteration index, i = 1,…,N p is number of particles index, j = 1,…,D is number of dimension index, X besti,j is local best position, i.e. it represents the best experience of individual particle, X lbesti is the neighborhood best position or the best experience of i th particle neighbors, c 1 and c 2 are two positive constants which represent the confidence for the experience of individual particle and neighbors, here c 1 = c 2 = 1, both r 1 and r 2 are two uniformly distributed random values between [0,1] that avoid falling into the local optimal, w is an inertial weight which represents the confidence of the previous velocity, here w is set to be uniformly distributed random values between [0,1]. To prevent particles from leaving the boundaries of search space, the velocity of particles is also clamped with a predefined maximum value V max . In

218815881

this way, each dimension of the search space is normalized to vary between [0, 1] and maximum velocity of particles in each dimension of search space is set to V max = 0.2. Therefore, the velocity vector of the saturated particles is randomized as: The neighborhood structure There are various neighborhood structures like star, ring, wheel and Von-Neumann (Engelbrecht, 2005) and Karimi, Pourtakdoust, and Nobahari (2011) have used a singlylinked ring structure for the neighborhoods. In singlylinked ring structure, as shown in Figure 1, a particle k has Figure 1. Singly-linked ring structure two neighbors, particle k-2 and particle k+1. In turn, particles k-2 and k+1 have not particle k as one of their neighbors. This topology keeps two neighbors for each particle, but breaks the mutual attraction between neighbors. In this way, the information is transmitted faster through the whole swarm than in the original ring topology. Quadratic interpolation operator Inspired by the mathematical fact that each three points produce a quadratic curve which it may have a minimum value in its valley, a quadratic interpolation operator was introduced by (Karimi, Pourtakdoust, & Nobahari, 2011). The quadratic interpolation operator utilizes the current experience of the intended particle, termed as main parent, and two other randomly chosen particles of the swarm. Subsequently, the fitness of the new particle is compared with the best experience of its main parent. The main parent's best position will be replaced by the new particle, if it is better. Note that, in this strategy, the position of the parent is not

218815881

changed and only its best position may be changed. According to the quadratic interpolation, the position of the new particle, generated by particle j and two random particles nb1 and nb2, at t-th iteration and j-th dimension of the search space is calculated as: (20) where, x i (t,j), x nb1 (t,j), x nb2 (t,j) and x new (t,j) indicate the j-th component of i-th, n b1 , n b2 and new particle position vectors at t-th iteration, respectively. Proposed heuristic estimation algorithms As previously mentioned, most researches for APE are based on the conventional strategy namely EAM. The EAM is a single step estimation strategy, directly estimate aerodynamic parameters. The EBM algorithm first proposed in (Mohammadi & Massoumnia, 2000) for APE in a two-step estimation strategy. In the 1 st step the aerodynamic forces and moments are estimated, then after passing through a designed smoothing filter, in 2 nd step aerodynamic parameters are estimated. In current research, the modified PSO algorithm, previously described, is combined with both EAM and EBM strategies, and new algorithms are developed. EAM-PSO algorithm In practice, the proposed EAM-PSO algorithm requires inertial navigation system measurement data including linear accelerations and angular rates in rolling body frame, , and control surface deflection angle for estimating control surface contribution in aerodynamic forces and moments. Therefore, in this research, after preparing the measurement data from 6DOF simulation model explained in section 4, an optimization problem is formed. The objective of this problem is to match measured values of [ , , , , ,

218815881

] m xm ym zm m m m a a a p q r = y with their estimated values [ , , , , , ] x y z a a a p q r By defining estimation state vector as , and estimated aerody namic para meters vector as x z z zq l lp m m mq , rolling airframe equations of motion, Eq. (2)-Eq. (10), can be stated in the following form: As can be seen, here, the estimation model is stated as a function of estimated aerodynamic coefficients. The estimated accelerations are calculated as: The EAM-PSO algorithm flow chart and pseudo code are shown in Figure 2 and Table 1, respectively. According to this algorithm, in initialization step, Aero X  is randomly selected and the estimation model of Eq. (24) is numerically integrated by 4 th order Runge Kutta method (Tan et al., 2012) and the estimated data vector is calculated. Note that numerical integration must be done each time that the objective function J EAM is calculated. This makes EAM-PSO to an algorithm having high computational time. During each iteration of estimation process, velocity and position vectors are updated. After that, particles personal and local best positions are calculated. Stopping criterion is set to be a predefined number of iterations. EBM-PSO algorithm In EBM-PSO, in addition to INS, an extra accelerometer set is also needed. The first accelerometers set is fixed at the position [R x1 , 0, 0] T from the C.G, and supplies the measured data Substituting

218815881

dp dt , dq dt and dr dt from Eq. (5)-Eq. (7) into Eq. (28) for both accelerometers yields (Larsson, 2013): (29) Or in compact form: Two steps of EBM-PSO are described in next subsections. EBM-PSO First Step In first step of EBM-PSO, the aerodynamic forces and should be estimated using Eq. (29). But the problem is that matrix A is not a full rank matrix (rank of A≠6). Rolling airframe symmetry characteristics causes this problem. Hence, by adding the relation of second row with that of fifth row, a new set of full rank equations will be found: EBM-PSO smoothing step The aerodynamic forces and moments calculated in first step are smoothed using Savitzky-Golay filter (Nirmal et al., 2016). In this filter a set of integers (A -n , A -(n-1) , …, A n-1 , A n ) are weighting coefficients to carry out the smoothing operation. The use of these weighting coefficients, known as convolution integers, turns out to be exactly equivalent to fitting the data to a polynomial, and it is computationally more effective and much faster. The smoothed data point by the Savitzky-Golay algorithm is given by the following equation: where, for a filter width (2n + 1) = 7 weighing coefficients are as follows: A -3 = -2, A -2 = 3, A -1 = 6, A 0 = 7, A 1 = 6, A 2 = 3, A 3 = -2. EBM-PSO second step In second step, the aerodynamic parameters are estimated by forming an optimization problem. x z

218815881

z zq l lp m m mq as aerodynamic parameters to be estimated (or optimization design variable), the estimation model can be summarized in the following vector form: The objective of this problem is to match calculated forces and moments with their estimated values, during the intended interval time, i = 1 … N, i.e.: Solving the optimization problem, all nine aerodynamic parameters Aero X  can be estimated. The EBM-PSO optimization algorithm showed in Figure 3. The pseudo code is also presented in Table 2. Noting that in first step F A and M are estimated before aerodynamic modeling in second step, so this strategy is named EBM. Advantage of EBM-PSO algorithm is that numerical integration is removed in it, and this provides a low computational run time. SEBM-PSO algorithm In separated EBM-PSO or SEBM-PSO algorithm, the estimation model of Eq. (33) is separated into the following set of equations: Considering the above set of equations, the nine aerodynamic coefficients can be estimated by solving the following four separate optimization problems: First is an optimization problem with one-dimensional design vector [ ] x C  , and its objective function is formulated as: Second optimization problem has three design variables [ , , ] z z zq C C C α δ    , with the objective function: Third, has a two dimensional design vector l lp C C   , with the objective function: And finally, fourth problem has a three dimensional design vector [ , , ] m m mq

218815881