-
Notifications
You must be signed in to change notification settings - Fork 5
/
carlasim.yaml
251 lines (224 loc) · 8.62 KB
/
carlasim.yaml
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
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
## Defines configurations of Carla simulation
## General setup
sim_duration: 600 # sec
# Carla recorder for simulation replay
carla_recorder: On
## World
world:
# Rendering must be on for semantic segmentation output
no_rendering: False
# Synchronous mode and the simulation time step (must be <= 0.1 sec for simulation fidelity)
sync_mode: True
# Note: For behavior agent's controller to work, time step 0.05 is recommended.
# If autopilot is used, then 0.1 is sufficient.
delta_seconds: &delta_seconds 0.1 # seconds per tick
# Built-in map
map: Town04
# Carla weathers (in alphabetic order)
# 0 - ClearNoon
# 1 - ClearSunset
# 2 - CloudyNoon
# 3 - CloudySunset
# 4 - Default
# 5 - HardRainNoon
# 6 - HardRainSunset
# 7 - MidRainSunset
# 8 - MidRainyNoon
# 9 - SoftRainNoon
# 10 - SoftRainSunset
# 11 - WetCloudyNoon
# 12 - WetCloudySunset
# 13 - WetNoon
# 14 - WetSunset
weather: 0
# Currently, 2 methods to control the ego vehicle are provided:
# 1. Autopilot:
# Use the traffic manager built in Carla. It allows for some basic behavior settings,
# but the car can only roam around randomly.
# For more info: https://github.com/carla-simulator/carla/issues/2966
# 2. Behavior Agent:
# Use the BehaviorAgent class defined in the "agent" package found in Carla's repository.
# This utility class is used in several example codes for demonstrations. However, it is not
# officially documented and its use here is solely based on the examination of the example codes
# with some improvements and fundamental bug fixes. I also extend the behavior agent such that
# the controller follows look ahead point placed a certain distance ahead instead of just the
# next waypoint.
# The following are configurations for the 2 mentioned methods.
## Autopilot (uncomment this section to use autopilot)
# autopilot:
# auto_lane_change: True # This lets the car to make lane changes if necessary, not really useful in our case
# ignore_lights_percentage: 100
# vehicle_percentage_speed_difference: 0 # When set to negative, vehicles move beyond the speed limit
## Behavior Agent (uncomment this section to use behavior agent)
behavior_agent:
ignore_traffic_light: True
behavior: 'normal' # 'cautious', 'normal', 'aggressive'
# x, y, z in Carla's coordinate system (left-handed z-up)
# These are just rough coordinates.
# The nearest carla.Waypoints will be retrieved around these rough goals.
goals:
-
x: 12
y: 100
z: 0
-
x: 8
y: 0
z: 0
# -
# x: 12
# y: 0
# z: 0
-
x: 8
y: -100
z: 0
-
x: 150
y: -170
z: 0
-
x: 255
y: -205
z: 0
## Ego Vehicle
ego_veh:
# If True, use a random spawn point and ignore the following spawn point settings
use_random_spawn_point: False
# x, y, z in Carla's coordinate system (left-handed z-up)
spawn_location:
- 16.5
- 180
- 0
# pitch, yaw, roll (deg) in Carla's coordinate system (right-handed z-down)
spawn_orientation:
- 0.0
- -90.0
- 0.0
# Carla uses center of gravity as the model's origin
# The following is roughly measured from Carla's Mustang model
# Carla's Mustang:
# CG to rear axle: 1.4
# CG to front bumper: 2.4
raxle_to_cg: &raxle_to_cg 1.4 # (m) Distance from rear axle to center of gravity in x
raxle_to_fbumper: &raxle_to_fbumper 3.8 # (m) Distance from rear axle to front bumper in x
## Sensors
sensor:
# Ref: https://carla.readthedocs.io/en/0.9.9/ref_sensors/
# Parameters for carla APIs have to be strings
gnss:
pos_x: -1.4 # Installation position x with respect to center of gravity
noise_alt_bias: '0.0' # Mean parameter in the noise model for altitude.
noise_alt_stddev: '0.0' # Standard deviation parameter in the noise model for altitude.
# 0.00001 deg in longitude and latitude corresponds roughly to 1.1 m in x and y in Carla's example maps.
# This has been tesed using Town03 and Town04.
noise_lat_bias: '0.0' # Mean parameter in the noise model for latitude.
noise_lat_stddev: '0.0' # Standard deviation parameter in the noise model for latitude.
noise_lon_bias: '0.0' # Mean parameter in the noise model for longitude.
noise_lon_stddev: '0.0' # Standard deviation parameter in the noise model for longitude.
imu:
pos_x: -1.4 # Installation position x with respect to center of gravity
noise_accel_stddev_x: '0.0' # Standard deviation parameter in the noise model for acceleration (X axis).
noise_accel_stddev_y: '0.0' # Standard deviation parameter in the noise model for acceleration (Y axis).
noise_accel_stddev_z: '0.0' # Standard deviation parameter in the noise model for acceleration (Z axis).
noise_gyro_bias_x: '0.0' # Mean parameter in the noise model for the gyroscope (X axis).
noise_gyro_bias_y: '0.0' # Mean parameter in the noise model for the gyroscope (Y axis).
noise_gyro_bias_z: '0.0' # Mean parameter in the noise model for the gyroscope (Z axis).
noise_gyro_stddev_x: '0.0' # Standard deviation parameter in the noise model for the gyroscope (X axis).
noise_gyro_stddev_y: '0.0' # Standard deviation parameter in the noise model for the gyroscope (Y axis).
noise_gyro_stddev_z: '0.0' # Standard deviation parameter in the noise model for the gyroscope (Z axis).
# For virtual odometry
# This is not a carla implementation so use float instead
noise_vx_bias: 0.0 # (m/s) Mean in the noise model for vx
noise_vy_bias: 0.0 # (m/s) Mean in the noise model for vx
noise_vx_stddev: 0.0 # (m/s) Standard deviation in the noise model for vx
noise_vy_stddev: 0.0 # (m/s) Standard deviation in the noise model for vy
# Virtual odometry uses velocities of ego vehicle's actor directly,
# which is found to lag behind Carla's IMU by 1 simulation step.
# To recover that, virtual odometry's velocities are added with acceleration times simulation step
delta_seconds: *delta_seconds # Simulation time step
front_camera:
# Semantic segmentation image
pos_x: 0.6 # Installation position x with respect to center of gravity
pos_z: 1.5 # Installation position z with respect to center of gravity (+: up)
res_h: '800' # (px) horizontal resolution
res_v: '600' # (px) vertical resolution
fov: '60' # (deg)
## Ground truth extraction
gt:
pose:
raxle_to_cg: *raxle_to_cg # (m) Distance from rear axle to center of gravity in x
raxle_to_fbumper: *raxle_to_fbumper # (m) Distance from rear axle to front bumper in x
lane:
radius: 10 # (m) Search radius for lane markings
## Recorder
# Each toggle specifies whether its corresponding data are recoreded when the recorder is activated
# Note the names must match the ones definded in its class
# Using names that don't exist causes KeyError during runtime
# Example format for sensor data:
# sensor:
# sensor_name: <- The sensor name (sensor's attribute)
# data_name_1: On/Off <- The data name (i.e. the key in the data buffer)
# Example format for ground truth data:
# gt:
# static:
# gt_name: On/Off <- The static ground truth type name (e.g. pose, lane)
# seq:
# gt_name: <- The sequential ground truth type name (e.g. pose, lane)
# data_name_1: On/Off <- The data name (i.e. the key in the data buffer)
recorder:
# Sensor data
sensor:
imu:
timestamp: On
# Velocities
vx: On
vy: On
# Angular velocities
gyro_x: Off
gyro_y: Off
gyro_z: On
# Accelerations
accel_x: Off
accel_y: Off
accel_z: Off
gnss:
timestamp: On
lat: Off
lon: Off
alt: Off
x: On
y: On
z: On
semantic_camera:
timestamp: On
ss_image: On
depth_camera:
timestamp: On
depth_buffer: On
# Ground truth
gt:
# Static
static:
traffic_sign: on
# Sequential
seq:
pose:
# Rear axle's pose
raxle_location: On
raxle_orientation: On
lane:
# The properties of the closest Carla.Waypoint
in_junction: On
lane_id: On
into_junction: On
# Carla.LaneMarking objects
left_marking: On
next_left_marking: On
right_marking: On
next_right_marking: On
# c0 and c1 coefficients
left_marking_coeffs: On
next_left_marking_coeffs: On
right_marking_coeffs: On
next_right_marking_coeffs: On