-
Notifications
You must be signed in to change notification settings - Fork 0
/
env.py
executable file
·181 lines (162 loc) · 7.1 KB
/
env.py
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
181
import pygame
import numpy as np
import math
import time
from gym.spaces.box import Box
class Reacher:
def __init__(self, screen_size=1000, link_lengths = [200, 140, 100], joint_angles=[0, 0, 0], target_pos = [869,430], render=False):
# Global variables
self.screen_size = screen_size
self.link_lengths = link_lengths
self.joint_angles = joint_angles
self.num_actions=3 # equals to number of joints - 1
self.num_observations= 2*(self.num_actions+2)
self.L = 8 # distance from target to get reward 2
self.action_space=Box(-100,100, [self.num_actions])
self.observation_space=Box(-1000,1000, [2*(self.num_actions+2)])
# The main entry point
self.render=render
if self.render == True:
self.screen = pygame.display.set_mode((self.screen_size, self.screen_size))
pygame.display.set_caption("Reacher")
else:
pass
self.is_running = 1
# self.target_pos=[self.screen_size/4, self.screen_size/4]
self.target_pos=target_pos
self.steps=0
self.max_episode_steps=50
self.near_goal_range=0.5
# Function to compute the transformation matrix between two frames
def compute_trans_mat(self, angle, length):
cos_theta = math.cos(math.radians(angle))
sin_theta = math.sin(math.radians(angle))
dx = -length * sin_theta
dy = length * cos_theta
T = np.array([[cos_theta, -sin_theta, dx], [sin_theta, cos_theta, dy], [0, 0, 1]])
return T
# Function to draw the current state of the world
def draw_current_state(self, ):
# First link in world coordinates
T_01 = self.compute_trans_mat(self.joint_angles[0], self.link_lengths[0])
origin_1 = np.dot(T_01, np.array([0, 0, 1]))
p0 = [0, 0]
p1 = [origin_1[0], -origin_1[1]] # the - is because the y-axis is opposite in world and image coordinates
# Second link in world coordinates
T_12 = self.compute_trans_mat(self.joint_angles[1], self.link_lengths[1])
origin_2 = np.dot(T_01, np.dot(T_12, np.array([0, 0, 1])))
p2 = [origin_2[0], -origin_2[1]] # the - is because the y-axis is opposite in world and image coordinates
# Third link in world coordinates
T_23 = self.compute_trans_mat(self.joint_angles[2], self.link_lengths[2])
origin_3 = np.dot(T_01, np.dot(T_12, np.dot(T_23, np.array([0, 0, 1]))))
p3 = [origin_3[0], -origin_3[1]] # the - is because the y-axis is opposite in world and image coordinates
# Compute the screen coordinates
# print(p0,p1,p2,p3)
p0_u = int(0.5 * self.screen_size + p0[0])
p0_v = int(0.5 * self.screen_size + p0[1])
p1_u = int(0.5 * self.screen_size + p1[0])
p1_v = int(0.5 * self.screen_size + p1[1])
p2_u = int(0.5 * self.screen_size + p2[0])
p2_v = int(0.5 * self.screen_size + p2[1])
p3_u = int(0.5 * self.screen_size + p3[0])
p3_v = int(0.5 * self.screen_size + p3[1])
# Draw
if self.render == True:
self.screen.fill((0, 0, 0))
pygame.draw.line(self.screen, (255, 255, 255), [p0_u, p0_v], [p1_u, p1_v], 5)
pygame.draw.line(self.screen, (255, 255, 255), [p1_u, p1_v], [p2_u, p2_v], 5)
pygame.draw.line(self.screen, (255, 255, 255), [p2_u, p2_v], [p3_u, p3_v], 5)
pygame.draw.circle(self.screen, (0, 255, 0), [p0_u, p0_v], 10)
pygame.draw.circle(self.screen, (0, 0, 255), [p1_u, p1_v], 10)
pygame.draw.circle(self.screen, (0, 0, 255), [p2_u, p2_v], 10)
pygame.draw.circle(self.screen, (255, 0, 0), [p3_u, p3_v], 10)
pygame.draw.circle(self.screen, (255, 255, 0), np.array(self.target_pos).astype(int), 10)
# Flip the display buffers to show the current rendering
pygame.display.flip()
else:
pass
return [p0_u,p0_v,p1_u,p1_v,p2_u,p2_v,p3_u,p3_v]
def reset(self,):
self.steps=0
self.joint_angles = np.array([0.1,0.1,0.1])*180.0/np.pi
if self.render == True:
self.screen = pygame.display.set_mode((self.screen_size, self.screen_size))
pygame.display.set_caption("Reacher")
else:
pass
self.is_running = 1
pos_set=self.draw_current_state()
# return np.array([np.concatenate((pos_set,self.link_lengths))])
return np.array([np.concatenate((pos_set,self.target_pos))]).reshape(-1)
def step(self,action):
# Get events and check if the user has closed the window
if self.render == True:
for event in pygame.event.get():
if event.type == pygame.QUIT:
self.is_running = 0
break
else:
pass
# Change the joint angles (the increment is in degrees)
change=np.random.uniform(-1,1,size=3)
# self.joint_angles[0] += 0.1
# self.joint_angles[1] += 0.2
# self.joint_angles[2] += 0.3
# self.joint_angles[0] += change[0]
# self.joint_angles[1] += change[1]
# self.joint_angles[2] += change[2]
# print(action)
self.joint_angles[0] += action[0]
self.joint_angles[1] += action[1]
self.joint_angles[2] += action[2]
# Draw the robot in its new state
# print(action)
pos_set=self.draw_current_state()
# if abs(pos_set[6]-self.target_pos[0])<self.L and abs(pos_set[7]-self.target_pos[1])<self.L:
# reward = 2
# else:
# reward = 0
# reward_0=1000
# reward = reward_0 * np.exp(-np.sqrt(abs(pos_set[6]-self.target_pos[0])**2+abs(pos_set[7]-self.target_pos[1])**2))
# print(reward) #e-100
reward_0=100.0
pos2goal_distance=np.sqrt(abs(pos_set[6]-self.target_pos[0])**2+abs(pos_set[7]-self.target_pos[1])**2)
reward = reward_0 / (pos2goal_distance+1)
# time.sleep(0.5)
self.steps+=1
if self.steps > self.max_episode_steps:
self.steps=0
return np.array([np.concatenate((pos_set,self.target_pos))]).reshape(-1), reward, True
else:
return np.array([np.concatenate((pos_set,self.target_pos))]).reshape(-1), reward, False
if __name__ == "__main__":
screen_size = 1000
# link_lengths = [200, 140, 100, 80]
link_lengths = [200, 140, 100]
joint_angles = [0, 0, 0, 0]
target_pos=[screen_size/4, screen_size/4]
reacher=Reacher(screen_size, link_lengths, joint_angles,target_pos)
# reacher.reset()
num_steps=50
# Loop until the window is closed
step=0
while reacher.is_running:
print(step)
step+=1
reacher.step()
if step >= num_steps:
reacher.is_running=0
# for step in range (num_steps):
# print(step)
# if reacher.is_running:
# reacher.step()
reacher.reset()
# print(reacher.is_running)
step=0
while reacher.is_running:
print(step)
step+=1
pos=reacher.step()
print(pos,len(pos))
if step >= num_steps:
reacher.is_running=0