-
Notifications
You must be signed in to change notification settings - Fork 45
/
config.py
57 lines (53 loc) · 2.07 KB
/
config.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
import rospkg
path = rospkg.RosPack().get_path('rl_nav')
# Environment related
maze_id = 1
deterministic = False
continuous = True # if the action space is continuous. True for PPO, DDPG and False for DQN
# image
input_dim = (48, 64, 3)
# reward parameter
r_arrive = 100
r_collision = -50 # -100
Cr = 100
Cp = -0.05 # time step penalty
# goal
goal_space = []
start_space = []
test_space = []
goal_space_nav0 = [[-2.8, -4.0]]
test_space_nav0 = [goal_space_nav0[-1]]
goal_space_nav1 = [[-3.8, -3.8], [-1.5, -4], [-3, -2.2]]
test_space_nav1 = [goal_space_nav1[-1]]
goal_space_nav2 = [[-1.8, -3.2], [-2, -2.6], [-3, -3.2], [-0.8, -3.5], [-0.4, -2.6], [-2.5, -3], [-3, -3], [-2, -1.9],
[-1.3, -3.8], [-3.3, -4], [-3, -3.8], [-1.2, -3.2]]
test_space_nav2 = [goal_space_nav2[-1], goal_space_nav2[-2]]
start_space_nav0 = [[1., -3.], [1., -4.], [0, -4.], [0, -1.8], [-1, -3], [-1.8, -1.9], [-1.7, -3.8]]
start_space_nav1 = [[-0.4, -2.6]] # ,4.0/3*math.pi
start_space_nav2 = [[0.65, -3.0]] # ,1.0/2*math.pi
start_space_nav3 = [[-3.8, -3.5], [-3.8, -4.1], [0, -3], [-1, -4.1], [-2, -4.1], [-3, -4.1], [-3, -2.5], [-1.2, -3.2],
[-1.2, -2.5], [-2, -3.2], [0, -2.5], [0, -3.5], [0, -4], [-2.5, -4.1], [-2.5, -3.2], [-3, -2.5],
[-3.5, -2.8], [-3.5, -2.5]]
goal_space.append(goal_space_nav0)
goal_space.append(goal_space_nav1)
goal_space.append(goal_space_nav2)
test_space.append(test_space_nav0)
test_space.append(test_space_nav1)
test_space.append(test_space_nav2)
start_space.append(start_space_nav0)
start_space.append(start_space_nav1)
start_space.append(start_space_nav2)
start_space.append(start_space_nav3)
Cd = 0.354/2 # = r when testing
# max linear velocity
v_max = 0.3 # 0.5 # m/s
# max angular velocity
w_max = 1.0 # 1.2 # rad/s
# VAE related
latent_vector_dim = 32
vae_weight = path + '/scripts/worldModels/models/vae_weights.h5'
# RNN related
dir_name = 'maze1_rnn_data' # which dir to save the rnn data
rnn_weight = path + '/scripts/worldModels/models/rnn_weights.h5'
episode = 2 # < 1000, for testing RNN
frame = 25 # < 50, for testing RNN