-
Notifications
You must be signed in to change notification settings - Fork 0
/
syncSetupExp.py
193 lines (147 loc) · 7.81 KB
/
syncSetupExp.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
182
183
184
185
186
187
188
189
190
191
import pickle
import carla
from carla import Rotation, Location
import pygame
import torch
from carla import ColorConverter
from carlaUtils import set_weather, set_sync, create_cam, to_data_in, to_salient_var, dummy_detector, retrieve_data, \
norm_salient_input, rollout_nll, Detector_Outputs, SimSnapshot
from customAgent import CustomAgent
from pems import load_model_det, PEMClass_Deterministic, PEMReg_Aleatoric
from render_utils import world_to_cam_viewport, depth_array_to_distances, get_image_as_array, draw_image
# class Sim_Snapshot:
# def __init__(self,
# time_step: int,
# model_det: bool,
# true_det: bool,
# true_centre: Optional[np.ndarray],
# model_centre: Optional[np.ndarray],
# cam_trans: carla.Transform,
# ego_v: carla.Vehicle,
# adv_v: carla.Vehicle):
#
# self.time_step = time_step
# self.model_det: model_det
# self.true_det = true_det
# if true_centre is not None:
# self.true_centre: Tuple[float, float] = tuple(true_centre)
# else:
# self.true_centre = None
#
# if model_centre is not None:
# self.model_centre: Tuple[float, float] = tuple(model_centre)
# self.cam_loc = to_loc_tuple(cam_trans)
# self.cam_rot = to_rot_tuple(cam_trans)
# self.ego_loc = to_loc_tuple(ego_v.get_transform())
# self.ego_rot = to_rot_tuple(ego_v.get_transform())
# self.adv_loc = to_loc_tuple(adv_v.get_transform())
# self.adv_rot = to_rot_tuple(adv_v.get_transform())
# class SnapshotEncoder(json.JSONEncoder):
# def default(self, obj):
# if isinstance(obj, Sim_Snapshot):
# return obj.__dict__
# return json.JSONEncoder.default(self, obj)
def run():
actor_list = []
try:
client = carla.Client('localhost', 2000)
client.set_timeout(10.0)
world = client.get_world()
# Load desired map
client.load_world("Town01")
set_sync(world, client, 0.05)
set_weather(world, 0, 0, 0, 0, 0, 75)
bpl = world.get_blueprint_library()
# Spawn the ego vehicle
ego_bp = bpl.find('vehicle.lincoln.mkz_2017')
ego_bp.set_attribute('role_name', 'ego')
ego_vehicle = world.spawn_actor(ego_bp, carla.Transform(Location(207, 133, 0.5), Rotation(0, 0, 0)))
world.tick()
actor_list.append(ego_vehicle)
# Load Perception model
pem_class = load_model_det(PEMClass_Deterministic(14, 1, use_cuda=False),
"models/det_baseline_full/pem_class_train_full")
pem_reg = load_model_det(PEMReg_Aleatoric(14, 2, use_cuda=False), "models/al_reg_full/pem_reg_al_full")
norm_stats = torch.load("models/norm_stats_mu.pt"), torch.load("models/norm_stats_std.pt")
n_func = lambda s_inputs, norm_dims: norm_salient_input(s_inputs, norm_stats[0], norm_stats[1], norm_dims)
# Create Cameras
cam_w, cam_h = 1242, 375
ego_cam, rgb_queue = create_cam(world, ego_vehicle, (cam_w, cam_h), 82, Location(2, 0, 1.76), Rotation())
depth_cam, depth_queue = create_cam(world, ego_vehicle, (cam_w, cam_h), 82, Location(2, 0, 1.76), Rotation(),
'depth')
# Spawn other vehicle
other_bp = bpl.find('vehicle.nissan.patrol')
ego_pos = ego_vehicle.get_location()
ego_forward = ego_vehicle.get_transform().get_forward_vector()
ego_right = ego_vehicle.get_transform().get_right_vector()
other_vehicle = world.spawn_actor(other_bp, carla.Transform(
ego_vehicle.get_location() + ego_forward * 20, ego_vehicle.get_transform().rotation))
world.tick()
actor_list.append(other_vehicle)
print(f'created {other_vehicle.type_id}')
# Set ego vehicle behaviour
# ego_vehicle.set_autopilot(True)
# agent = BasicAgent(ego_vehicle)
agent = CustomAgent(ego_vehicle)
other_vehicle.set_autopilot(True)
spectator = world.get_spectator()
pygame.init()
# py_display = pygame.display.set_mode((cam_w, cam_h * 2), pygame.HWSURFACE | pygame.DOUBLEBUF)
py_display = pygame.display.set_mode((cam_w, cam_h), pygame.HWSURFACE | pygame.DOUBLEBUF)
lights_list = world.get_actors().filter("*traffic_light*")
adv_v = world.get_actors().filter("vehicle.nissan.patrol")[0]
for l in lights_list:
l.set_red_time(100)
rollout_log = []
for i in range(350):
w_frame = world.tick()
# Follow ego on server window
spectator.set_transform(
carla.Transform(ego_vehicle.get_transform().location + Location(z=30), Rotation(pitch=-90)))
# Render sensor output
data_timeout = 2.0
current_rgb = retrieve_data(rgb_queue, w_frame, data_timeout)
current_depth = retrieve_data(depth_queue, w_frame, data_timeout)
# print(current_rgb.frame, current_depth.frame, w_frame)
distance_array = depth_array_to_distances(get_image_as_array(current_depth))
current_depth.convert(ColorConverter.LogarithmicDepth)
depth_im_array = get_image_as_array(current_depth)
draw_image(py_display, get_image_as_array(current_rgb))
# draw_image(py_display, depth_im_array, offset=(0, cam_h))
d_in = to_data_in(ego_cam.get_transform(), ego_cam.attributes, adv_v)
salient_vars = to_salient_var(d_in, n_func)
tru_adv_vp = world_to_cam_viewport(ego_cam.get_transform(), ego_cam.attributes,
adv_v.get_location() + Location(0, 0,
adv_v.bounding_box.extent.z)).astype(int)
# detection, cam_centroid, obstacle_depth = dummy_detector(salient_vars, adv_v, ego_cam, distance_array, 0.9)
# dummy_det, dummy_centroid, dummy_depth = dummy_detector(salient_vars, adv_v, ego_cam, distance_array, 0.9)
# m_detection, m_centroid, m_depth = model_detector(salient_vars, adv_v, ego_cam, distance_array, pem_class,
# pem_reg)
m_detection, m_centroid, m_depth = dummy_detector(salient_vars, adv_v, ego_cam, distance_array, 1.0)
d_outs = Detector_Outputs(tuple(tru_adv_vp),
tuple(m_centroid) if m_centroid is not None else None,
True,
m_detection)
pygame.draw.circle(py_display, (0, 255, 0), (tru_adv_vp[0], tru_adv_vp[1]), 5.0)
rollout_log.append(SimSnapshot(w_frame, d_in, d_outs))
if m_detection:
# print("BB-dist: \t", np.min([av.distance(ev) for av in adv_bb_verts for ev in ego_bb_verts]))
pygame.draw.circle(py_display, (255, 0, 0), (m_centroid[0], m_centroid[1]), 5.0)
# pygame.draw.rect(py_display, (255, 0, 0), pygame.Rect(cam_centroid[0] - 5, cam_h + cam_centroid[1] - 5, 10, 10), 2)
pygame.display.flip()
ego_vehicle.apply_control(agent.run_step(m_centroid, m_depth))
with open("data_outs/rollout_log.pickle", 'wb') as f:
pickle.dump(rollout_log, f)
with open("data_outs/rollout_log.pickle", 'rb') as f:
loaded_roll = pickle.load(f)
print(loaded_roll)
nll = rollout_nll(rollout_log, pem_class, pem_reg, n_func)
finally:
ego_cam.destroy()
depth_cam.destroy()
print("Actors to destroy: ", actor_list)
client.apply_batch([carla.command.DestroyActor(x) for x in actor_list])
pygame.quit()
print("Done")
if __name__ == "__main__":
run()