6.4. Test the pipeline for driving a mission

6.4.1. Prerequisites

6.4.1.1. Rosbag

  1. Download (see How the drive is structured for data from and for test days) and prepare (migrate, see Migrating Rosbags for pipeline testing purposes) a rosbag containing the recording of a driven trackdrive. Place the rosbag in the rosbags directory (as_ros/rosbags)

    1. Trackdrive and Autocross: We currently recommend rosbags recorded with Eva in May ‘22 (e.g. Run 370), see for migration here: Data recorded with Eva in May ‘22

    2. Skidpad: We currently recommend rosbags recorded with Eva in May ‘22 (e.g. Run 377), see for migration here: Data recorded with Eva in May ‘22

    3. EBS Test and Acceleration: We currently recommend rosbags recorded with Emma in KW 40 ‘22 (e.g. Run 592), see for migration here: Data recorded with Emma in KW 44 ‘22

6.4.1.2. Configuration

The autonomous system has a lot of parameters that can be changed and tuned. The next section will provide a small overview which parameters should be checked anyways.

Some paramaters regarding the dynamics of the vehicle have to be adapted according to meausrements inside the rosbag since e.g. the model predict control is going crazy if the drives faster than the trajectory is planned.

  1. Check the src/utilities/config/test_track.yaml config, especially

    1. Set the test_day to <test_day_name>_sim

    2. Set the track_layout according to the current mission. You might use the same layout for Trackdrive and Autocross since they might use the same track layout.

  2. Check the src/camera_perception/config/perception.yaml, especially

    1. Set camera_usage_strategy according to how you want to use the different cameras.

    2. Set detection_image to whether a detection image should be created or not. Keep in mind you can create a detection image by using the rosbag migrator in post processing: Creating Visualization for recorded Rosbags

  3. Check the src/slam/config/slam.yaml, especially

    The most important parameters and their meaning are described in the doc string of slam

    1. Set preloading/activated according to whether you want to preload a manual created map or previously tracked map to preload into slam

    2. Set preloading/align according to whether you want to align the preloaded map

    3. Set <sensor_name>/active according to whether you want to and can use this sensor, e.g. maybe one wheelspeed sensor was defect.

  4. Check the src/path_planning/config/path_planning.yaml, especially

    1. Set track_width_min according to you measured minimum track width

  5. Check the src/motion_planning/config/local_motion_planning.yaml, especially

    1. Set velocity_profile/a_x_factor according to which measurements you can observe inside the rosbag.

    2. Set velocity_profile/a_y_factor according to which measurements you can observe inside the rosbag.

    3. Set velocity_profile/v_max according to which measurements you can observe inside the rosbag.

    4. Set velocity_profile/a_x_factor_completed according to which measurements you can observe inside the rosbag.

  6. Check the src/motion_planning/config/global_motion_planning.yaml, especially

    1. Set velocity_profile/a_x_factor according to which measurements you can observe inside the rosbag.

    2. Set velocity_profile/a_y_factor according to which measurements you can observe inside the rosbag.

    3. Set velocity_profile/v_max according to which measurements you can observe inside the rosbag.

    4. Set velocity_profile/a_x_factor_completed according to which measurements you can observe inside the rosbag.

6.4.1.3. Map Generation

If you want to use preloading you might need to generate a manual map first. You there are 3 utilities you can use

  1. Generating a rule-defined Acceleration map

  2. Generating a rule-defined Skidpad map

  3. Creating Tracks from drone footage

6.4.2. Running the pipeline

6.4.2.1. General Approach

There exists a custom launch file to debug the complete pipeline (debug_pipeline.launch in utilities ros package). Among others, you can specify to not run specific ROS modules. See below.

roslaunch utilities debug_pipeline.launch input_rosbag:=/workspace/as_ros/rosbags/<rosbag_name>.bag mission:=<mission_id> vehicle:=<vehicle_name>
Table 6.1 Mission IDs

Mission ID

Mission Name

0

Manual Driving

1

Acceleration

2

Skidpad

3

Trackdrive

4

Braketest / EBS Test

5

Inspection

6

Autocross
Table 6.2 Vehicle Names

Vehicle

Vehicle Name (ROS Launch Parameter)

CM-21x

eva

CM-22x

emma

You may want to adapt the command according to your needs:

  1. If you use a rosbag with compressed images, you must append republish:=yes.

  2. If you do not need or want to execute a specific as ros module, you can exclude the module by appending the following: <module_name>:=no, e.g. control:=no. Possible values: perception, mapping, slam, path_planning, motion_planning, control.

  3. If you want to profile the AS ROS modules, you can apppend: profiling:=yes.

  4. If you need to use an older camera matrix, you can specify the matrices subdirectory by appending: matrices_subdirectory:=<matrices_subdirectory>. Parameter value does not need a / at the beginning or end of the path part.

  5. If you want to log the cpu usage of the various ros nodes, you can append cpu_monitor:=yes.

  6. If you want slam to send additional debug informations like the current predicted and actual measurements, you can append send_debug_informations:=yes.

  7. If your rosbag already includes tf transforms, you can append transforms:=no (NOT RECOMMENDED).

  8. If you want to use the camera perception, but your input rosbag does not contain uncompressed images, you need to uncompress by either

Table 6.3 Existing camera matrices subdirectories

Vehicle

Time Frame

Runs

Camera matrices subdirectory

Comment

Eva

May 2022

368-388

eva_may

All runs

Emma

KW40 (08. Oct 22)

506-520

emma_kw40_accel

Acceleration Runs

Emma

KW40 (08. Oct 22)

527-533

emma_kw40_skidpad

Skidpad Runs

Emma

KW44 (05. - 06. Nov 22)

584-611

emma_kw44_accel

Acceleration Runs

Emma

KW44 (05. - 06. Nov 22)

612-614

emma_kw44_skidpad

Skidpad Runs

Emma

KW05 (05. - 05. Feb 23)

621-629

emma_kw05

Acceleration/Skidpad Runs

Rennate

KW30 (27. Jul 23)

692-700

rennate_july

Brake Test

Rennate

KW43 (23. Oct 23)

n/a

validation_day

Camera Validation Day

Rennate

Mar - Apr 24

n/a

sensorfusion_720/_2k

Sensorfusion Testing

6.4.2.2. Current Recommendations

There exists recommendations for the simulating different disciplines. You need to download (see How the drive is structured for data from and for test days) and migrate (see Migrating Rosbags) the rosbags first.

6.4.2.2.1. Acceleration

roslaunch utilities debug_pipeline.launch input_rosbag:=/workspace/as_ros/rosbags/testing/run_592_m.bag mission:=1 vehicle:=emma matrices_subdirectory:=emma_kw44 cpu_monitor:=yes send_debug_informations:=yes

6.4.2.2.2. EBS Test

roslaunch utilities debug_pipeline.launch input_rosbag:=/workspace/as_ros/rosbags/testing/run_592_m.bag mission:=4 vehicle:=emma matrices_subdirectory:=emma_kw44 cpu_monitor:=yes send_debug_informations:=yes

6.4.2.2.3. Skidpad

roslaunch utilities debug_pipeline.launch input_rosbag:=/workspace/as_ros/rosbags/testing/run_377_m.bag mission:=2 vehicle:=eva matrices_subdirectory:=eva_may cpu_monitor:=yes send_debug_informations:=yes

6.4.2.2.4. Autocross

roslaunch utilities debug_pipeline.launch input_rosbag:=/workspace/as_ros/rosbags/testing/run_370_m.bag mission:=6 vehicle:=eva matrices_subdirectory:=eva_may cpu_monitor:=yes send_debug_informations:=yes

6.4.2.2.5. Trackdrive

roslaunch utilities debug_pipeline.launch input_rosbag:=/workspace/as_ros/rosbags/testing/run_370_m.bag mission:=3 vehicle:=eva matrices_subdirectory:=eva_may cpu_monitor:=yes send_debug_informations:=yes

6.4.2.3. Post Visualizing

See Creating Visualization for recorded Rosbags.