Implement separate parking controller

[a-0-dev]

• Input: EnvironmentModel.msg, ParkingInProgress.srv
• Output: TrajectoryMetaInput.msg, ParkingInProgress.srv Features:
• [x] Process environment data, measure parking spot sizes
• [x] Detect/Flag possible parking spots
• [ ] Perform complete parking process (pulling in, reporting success, pulling out again)
• [ ] Parallel parking, Perpendicular parking or both

[YuxuanSnow]

Wiki entry with additional information (to be expanded)

[YuxuanSnow]

Message prototype from uavcan for vehicle control

[YuxuanSnow]

Feedback message with current vehicle state

[YuxuanSnow]

Environment model message type

[YuxuanSnow]

Obstacle message type in environment model message

[YuxuanSnow]

Lane message type in obstacle

[YuxuanSnow]

Light switch message prototype from electronics (starts measuring when the switch detects the beginning of a parking spot (infinite distance) and sends out the message 5 times when a distance is detected again, which means the spot has ended, message contains length of the spot and the orthogonal distance)

[YuxuanSnow]

Barred area marker in env model (to be updated, currently uses old message)

enumeration of types is here (only MARKING_PARKING_SPOT_BLOCKED is relevant for you 209 )

[YuxuanSnow]

Currently, the node does not update the parking state as the function findPotentialSpots is not called here. This means that thet parking state machine is never entered (after line 167).

Also, the service handler is completely wrong -> it sets request and response at the same time, which is wrong, it should send request only and wait for acknowledgement from the Behaviour Tree.

[YuxuanSnow]

@MaslinuPoimal will try to make a simulation env with parking set up available until next week for live testing

[YuxuanSnow]

back parking algorithm already implemented. The problem is, we still have to subscribe successfully to have the input of our controller. Today we have tried to subscribe some rostopic to test the communication. @Priyanka-tig

[YuxuanSnow]

Lidar funktion (physical object detection) integrates now in our control program. We have found a ros bag, which includes the Lidar data. We will try to subscribe this message and test our program on the vehicle in the Lab on sunday at 11. @Priyanka-tig @Zeju1997

[YuxuanSnow]

@Zeju1997 The ROS will execute the function in the ROS::subscriber once a message occurred. There is no loop in the main function, but the message is subscribed continuously. For this reason, the function is executed continuously just like in a loop. We tried today to detect the slot length by use of the lidar data. We have written it in the sensor_data_callback, so that the detected length can be printed once a lidar message comes.

[YuxuanSnow]

Blick on the drive_ros_config, trajectory_tracking_controller, there are some examples on how to deal with the message with/in the uavcan.

[YuxuanSnow]

In the launch file(what we supposed to have), the current vehicle state we want to read is /canros/msg/phoenix_msgs/DriveState. In addition to last comment: the trajectory_tracking_control is the controller which changes the velocity and steering angle. In our parking process we can use its message.