|
extends Node3D |
|
class_name AIController |
|
|
|
var _player : Player |
|
|
|
# ------------------ Godot RL Agents Logic ------------------------------------# |
|
var heuristic := "human" |
|
var done := false |
|
# example actions |
|
|
|
var movement_action := Vector2(0.0, 0.0) |
|
var look_action := Vector2(0.0, 0.0) |
|
var jump_action := false |
|
var shoot_action := false |
|
|
|
var needs_reset := false |
|
var reward := 0.0 |
|
var n_steps_without_positive_reward = 0 |
|
var n_steps = 0 |
|
|
|
@onready var wide_raycast_sensor = $WideRaycastSensor |
|
@onready var narrow_raycast_sensor = $NarrowRaycastSensor |
|
|
|
func init(player): |
|
_player=player |
|
|
|
func set_team(value): |
|
wide_raycast_sensor.team = value |
|
narrow_raycast_sensor.team = value |
|
if value == 0: |
|
wide_raycast_sensor.team_collision_mask = 8 |
|
wide_raycast_sensor.enemy_collision_mask = 16 |
|
narrow_raycast_sensor.team_collision_mask = 8 |
|
narrow_raycast_sensor.enemy_collision_mask = 16 |
|
elif value == 1: |
|
wide_raycast_sensor.team_collision_mask = 16 |
|
wide_raycast_sensor.enemy_collision_mask = 8 |
|
narrow_raycast_sensor.team_collision_mask = 16 |
|
narrow_raycast_sensor.enemy_collision_mask = 8 |
|
|
|
|
|
func reset(): |
|
n_steps_without_positive_reward = 0 |
|
n_steps = 0 |
|
|
|
func reset_if_done(): |
|
if done: |
|
reset() |
|
|
|
func get_obs(): |
|
var obs = [] |
|
obs.append_array(wide_raycast_sensor.get_observation()) |
|
obs.append_array(narrow_raycast_sensor.get_observation()) |
|
return { |
|
"obs":obs |
|
} |
|
|
|
func get_reward(): |
|
var total_reward = reward + shaping_reward() |
|
if total_reward <= 0.0: |
|
n_steps_without_positive_reward += 1 |
|
else: |
|
n_steps_without_positive_reward -= 1 |
|
n_steps_without_positive_reward = max(0, n_steps_without_positive_reward) |
|
return total_reward |
|
|
|
func zero_reward(): |
|
reward = 0.0 |
|
|
|
func shaping_reward(): |
|
var s_reward = 0.0 |
|
return s_reward |
|
|
|
|
|
func set_heuristic(h): |
|
# sets the heuristic from "human" or "model" nothing to change here |
|
heuristic = h |
|
|
|
func get_obs_space(): |
|
var obs = get_obs() |
|
return { |
|
"obs": { |
|
"size": [len(obs["obs"])], |
|
"space": "box" |
|
}, |
|
} |
|
|
|
func get_action_space(): |
|
return { |
|
"movement_action" : { |
|
"size": 2, |
|
"action_type": "continuous" |
|
}, |
|
"look_action" : { |
|
"size": 2, |
|
"action_type": "continuous" |
|
}, |
|
"jump_action" : { |
|
"size": 2, |
|
"action_type": "discrete" |
|
}, |
|
"shoot_action" : { |
|
"size": 2, |
|
"action_type": "discrete" |
|
}, |
|
} |
|
|
|
func get_done(): |
|
return done |
|
|
|
func set_done_false(): |
|
done = false |
|
|
|
func set_action(action): |
|
movement_action = Vector2(clamp(action["movement_action"][0],-1.0,1.0), clamp(action["movement_action"][1],-1.0,1.0)) |
|
look_action = Vector2(clamp(action["look_action"][0],-1.0,1.0), clamp(action["look_action"][1],-1.0,1.0)) |
|
jump_action = action["jump_action"] == 1 |
|
shoot_action = action["shoot_action"] == 1 |
|
|
|
|
|
func _physics_process(delta): |
|
n_steps += 1 |
|
if n_steps > 4000: |
|
_player.needs_respawn = true |
|
|