File size: 6,479 Bytes
742ef47
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
import numpy as np
import pybullet as p
from surrol.gym.surrol_env import SurRoLEnv, RENDER_HEIGHT
from surrol.robots.ecm import Ecm
from surrol.utils.pybullet_utils import (
    get_link_pose,
    reset_camera
)
from surrol.utils.robotics import get_pose_2d_from_matrix


def goal_distance(goal_a, goal_b):
    assert goal_a.shape == goal_b.shape
    return np.linalg.norm(goal_a - goal_b, axis=-1)


class EcmEnv(SurRoLEnv):
    """
    Single arm env using ECM (active_track is not a GoalEnv)
    Refer to Gym fetch
    https://github.com/openai/gym/blob/master/gym/envs/robotics/fetch_env.py
    ravens
    https://github.com/google-research/ravens/blob/master/ravens/environments/environment.py
    """
    ACTION_SIZE = 3  # (dx, dy, dz) or cVc or droll (1)
    ACTION_MODE = 'cVc'
    DISTANCE_THRESHOLD = 0.005
    POSE_ECM = ((0.15, 0.0, 0.7524), (0, 30 / 180 * np.pi, 0))
    QPOS_ECM = (0, 0.6, 0.04, 0)
    WORKSPACE_LIMITS = ((0.45, 0.55), (-0.05, 0.05), (0.60, 0.70))
    SCALING = 1.

    def __init__(self, render_mode: str = None, cid = -1):
        # workspace
        self.workspace_limits = np.asarray(self.WORKSPACE_LIMITS)
        self.workspace_limits *= self.SCALING

        # camera
        self.use_camera = False

        # has_object
        self.has_object = False
        self.obj_id = None

        super(EcmEnv, self).__init__(render_mode, cid)

        # change duration
        self._duration = 0.1

        # distance_threshold
        self.distance_threshold = self.DISTANCE_THRESHOLD * self.SCALING

        # render related setting
        self._view_matrix = p.computeViewMatrixFromYawPitchRoll(
            cameraTargetPosition=(0.27 * self.SCALING, -0.20 * self.SCALING, 0.55 * self.SCALING),
            distance=1.80 * self.SCALING,
            yaw=150,
            pitch=-30,
            roll=0,
            upAxisIndex=2
        )

    def render(self, mode='rgb_array'):
        # TODO: check how to disable specular color when using EGL renderer
        if mode != "rgb_array":
            return np.array([])
        rgb_array = super().render(mode)
        if self.use_camera:
            rgb_cam, _ = self.ecm.render_image(RENDER_HEIGHT, RENDER_HEIGHT)
            rgb_array = np.concatenate((rgb_array, rgb_cam), axis=1)
        return rgb_array

    def compute_reward(self, achieved_goal: np.ndarray, desired_goal: np.ndarray, info):
        """ Sparse reward. """
        # d = goal_distance(achieved_goal, desired_goal)
        # return - (d > self.distance_threshold).astype(np.float32)
        return self._is_success(achieved_goal, desired_goal).astype(np.float32) - 1.

    def _env_setup(self):
        assert self.ACTION_MODE in ('cVc', 'dmove', 'droll')
        # camera
        if self._render_mode == 'human':
            reset_camera(yaw=150.0, pitch=-30.0, dist=1.50 * self.SCALING,
                         target=(0.27 * self.SCALING, -0.20 * self.SCALING, 0.55 * self.SCALING))

        # robot
        self.ecm = Ecm(self.POSE_ECM[0], p.getQuaternionFromEuler(self.POSE_ECM[1]),
                       scaling=self.SCALING)

        pass  # need to implement based on every task
        # self.obj_ids

    def _get_robot_state(self) -> np.ndarray:
        # TODO
        # robot state: eef pose in the ECM coordinate
        pose_rcm = get_pose_2d_from_matrix(self.ecm.get_current_position())
        return np.concatenate([
            np.array(pose_rcm[0]), np.array(p.getEulerFromQuaternion(pose_rcm[1]))
        ])

    def _get_obs(self) -> dict:
        robot_state = self._get_robot_state()
        # may need to modify
        if self.has_object:
            pos, _ = get_link_pose(self.obj_id, -1)
            object_pos = np.array(pos)
        else:
            object_pos = np.zeros(0)

        if self.has_object:
            achieved_goal = object_pos.copy()
        else:
            achieved_goal = np.array(get_link_pose(self.ecm.body, self.ecm.EEF_LINK_INDEX)[0])  # eef position

        observation = np.concatenate([
            robot_state, object_pos.ravel(),  # achieved_goal.copy(),
        ])
        obs = {
            'observation': observation.copy(),
            'achieved_goal': achieved_goal.copy(),
            'desired_goal': self.goal.copy()
        }
        return obs

    def _set_action(self, action: np.ndarray):
        """
        delta_position (3) and delta_theta (1); in world coordinate
        """
        # print('action: ', action)
        assert len(action) == self.ACTION_SIZE
        action = action.copy()  # ensure that we don't change the action outside of this scope
        if self.ACTION_MODE == 'cVc':
            # hyper-parameters are sensitive; need to tune
            # if np.linalg.norm(action) > 1:
            #     action /= np.linalg.norm(action)
            action *= 0.01 * self.SCALING  # velocity (HeadPitch, HeadYaw), limit maximum change in velocity
            dq = 0.05 * self.ecm.cVc_to_dq(action)  # scaled
            result = self.ecm.dmove_joint(dq)
            if result is False:
                return False
            else:
                return True
        elif self.ACTION_MODE == 'dmove':
            # Incremental motion in cartesian space in the base frame
            action *= 0.01 * self.SCALING
            pose_rcm = self.ecm.get_current_position()
            pose_rcm[:3, 3] += action
            pos, _ = self.ecm.pose_rcm2world(pose_rcm, 'tuple')
            joint_positions = self.ecm.inverse_kinematics((pos, None), self.ecm.EEF_LINK_INDEX)  # do not consider orn
            self.ecm.move_joint(joint_positions[:self.ecm.DoF])
        elif self.ACTION_MODE == 'droll':
            # change the roll angle
            action *= np.deg2rad(3)
            self.ecm.dmove_joint_one(action[0], 3)
        else:
            raise NotImplementedError

    def _is_success(self, achieved_goal, desired_goal):
        """ Indicates whether or not the achieved goal successfully achieved the desired goal.
        """
        d = goal_distance(achieved_goal, desired_goal)
        return (d < self.distance_threshold).astype(np.float32)

    def _sample_goal(self) -> np.ndarray:
        """ Samples a new goal and returns it.
        """
        raise NotImplementedError

    @property
    def action_size(self):
        return self.ACTION_SIZE

    def get_oracle_action(self, obs) -> np.ndarray:
        """
        Define a scripted oracle strategy
        """
        raise NotImplementedError