官网:https://carla.readthedocs.io/projects/ros-bridge/en/latest/carla_ackermann_control/
参考:https://blog.csdn.net/dfman1978/article/details/125077551
carla_ackermann_control
包使用 Ackermann messages 控制 CARLA
车辆。该包将 Ackermann messages 转换为CarlaEgoVehicleControl 消息。
它从 CARLA
读取车辆信息,并将该信息传递给一个基于 Python
的PID
控制器,称为 simple-pid
,以控制加速度和速度。
在 ros_bridge
中,除了使用 manual_control
包来直接使用键盘来控制仿真小车,还可以通过ackermann
(阿克曼)来控制。在 carla
和 autoware
联合仿真的过程中,autoware
最终的控制信息是要转换为 ackermann
控制指令,最后控制仿真车。
在使用 ROS 1
和 ROS 2
时,可以在 配置文件 中初始设置参数,在运行时通过 ROS 1
动态重新配置参数 dynamic_reconfigure。
通过 topic /carla/
向小车发送命令来测试设置。例如,运行如下命令,将角色名为 ego_vehicle
的自我车辆以10米/秒的速度向前移动:
rostopic pub /carla/ego_vehicle/ackermann_cmd ackermann_msgs/AckermannDrive \
"{steering_angle: 0.0, steering_angle_velocity: 0.0, speed: 10, acceleration: 0.0, jerk: 0.0}" -r 10
或者让车辆以1.22弧度的角度转弯时向前移动:
rostopic pub /carla/ego_vehicle/ackermann_cmd ackermann_msgs/AckermannDrive \
"{steering_angle: 1.22, steering_angle_velocity: 0.0, speed: 10, acceleration: 0.0, jerk: 0.0}" -r 10
上面的命令执行后,通过 rostopic
命令查看,该 topic
在没有订阅者的情况下车辆是不会运动的。
wohu@wohu-pc:~/tool$ rostopic info /carla/ego_vehicle/ackermann_cmd
Type: ackermann_msgs/AckermannDrive
Publishers:
* /rostopic_16525_1657090091755 (http://wohu-pc:39319/)
Subscribers: None
wohu@wohu-pc:~/tool$
其中 /carla/ego_vehicle/ackermann_cmd
就是发过来的 ackermann
命令,节点对这个命令进行解析,然后转换成 ros_bridge
节点可以识别的消息格式:/carla/ego_vehicle/vehicle_conttrol_cmd
,发给 /carla_ros_bridge
节点。
可以在 carla-ros-bridge/catkin_ws/src/ros-bridge/carla_ros_bridge/src/carla_ros_bridge/ego_vehicle.py
代码里看到
self.control_subscriber = node.new_subscription(
CarlaEgoVehicleControl,
self.get_topic_prefix() + "/vehicle_control_cmd",
lambda data: self.control_command_updated(data, manual_override=False),
qos_profile=10)
self.manual_control_subscriber = node.new_subscription(
CarlaEgoVehicleControl,
self.get_topic_prefix() + "/vehicle_control_cmd_manual",
lambda data: self.control_command_updated(data, manual_override=True),
qos_profile=10)
真正的执行函数是 control_command_updated
,其内容如下:
def control_command_updated(self, ros_vehicle_control, manual_override):
"""
Receive a CarlaEgoVehicleControl msg and send to CARLA
This function gets called whenever a ROS CarlaEgoVehicleControl is received.
If the mode is valid (either normal or manual), the received ROS message is
converted into carla.VehicleControl command and sent to CARLA.
This bridge is not responsible for any restrictions on velocity or steering.
It's just forwarding the ROS input to CARLA
:param manual_override: manually override the vehicle control command
:param ros_vehicle_control: current vehicle control input received via ROS
:type ros_vehicle_control: carla_msgs.msg.CarlaEgoVehicleControl
:return:
"""
if manual_override == self.vehicle_control_override:
vehicle_control = VehicleControl()
vehicle_control.hand_brake = ros_vehicle_control.hand_brake
vehicle_control.brake = ros_vehicle_control.brake
vehicle_control.steer = ros_vehicle_control.steer
vehicle_control.throttle = ros_vehicle_control.throttle
vehicle_control.reverse = ros_vehicle_control.reverse
vehicle_control.manual_gear_shift = ros_vehicle_control.manual_gear_shift
vehicle_control.gear = ros_vehicle_control.gear
self.carla_actor.apply_control(vehicle_control)
self._vehicle_control_applied_callback(self.get_id())
注释很清楚的说明了 接收 CarlaEgoVehicleControl
消息,发送给 CARLA
。
self.carla_actor.apply_control(vehicle_control)
这个函数就是来控制仿真车辆。
该包将 Ackermann messages 转换为CarlaEgoVehicleControl
消息的,实际在 ROS Bridge 中使用的是 CarlaEgoVehicleControl
消息发送给 CARLA
的,所以如果是 ROS
与 Autoware
对接那么就应该使用这个包,如果是和其它系统对接,比如 Apollo
,那么只需要将 Apollo
的相应消息转化为 CarlaEgoVehicleControl
格式即可。
Topic | Type | Description |
---|---|---|
/carla//ackermann_cmd | ackermann_msgs.AckermannDrive | Subscriber for steering commands |
Topic | Type | Description |
---|---|---|
/carla//ackermann_control/control_info | carla_ackermann_control.EgoVehicleControlInfo | The current values used within the controller (useful for debugging) |