Parking + Perception Integration. cleaned. ready to merge

[AadarshHegde123]

Integration of Perception's cone detection into parking team's parking space detection. We have also integrated this perception pipeline to work within the overall integration scheme outlined by the Planning team.

Video attached showcasing this integration's functionality. For reference, the ParkingSim component is the component that runs Parking team's overall state machine logic (ie. starting in Scanning mode, checking if we found a goal parking space and updating state to Parking). The RoutePlanningComponent is the component that will actually run planning logic depending on what state we are in.

The overall component pipeline is ConeDetector3D (Perception Team) -> ParkingSpotsDetector3D (Parking Team) -> ParkingSim (Parking Team) -> RoutePlanningComponent (Shared with other Teams)

The actual integrated planning logic is not included in this PR.

perception_integration_functionality.mp4

[KenC1014]

Constants are now moved under utils/constants.py

[krishauser]

Perception PR is not yet merged pending review. Wait for their updates.

[AadarshHegde123]

Perception has converted cones to be obstacles and updated their PR. This has been integrated and tested on this branch.

[AadarshHegde123]

Decided "middle ground" approach for calibration files is that calibration team will have a script that can convert their yaml files to the format perception and parking wants with cameras.yaml. With this approach parking and perception PRs can remain unchanged and are both functional.

[sonarqubecloud]

Quality Gate passed

Issues
96 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[krishauser]

Still awaiting resolution of mission_plan.py and utils.py comments.

[krishauser]

Trying to submit file specific comments again…

[KenC1014]

The newest commit that address the above comments includes the follows:

1. Removed lidar to vehicle transform from "constants.py". Instead, T_l2v was added to "camera.yaml", parking _detection component now reads in yaml file supplied by calibration as input arguments.
2. "cone_detection.py" was using T_l2v from "constants.py" before, now it reads this value from the "camera.yaml" instead. (perception code modified)
3. "parking_detection.yaml" was modified to take in proper arguments.
4. There were multiple definitions of "obstacle_detection" component in the computational graph. The duplication was removed.
5. vehicle length and width are now getting values from "gem_e4_geometry.yaml"

Flow was tested again with Rosbag after the above changes were made.

[AadarshHegde123]

Updated to include some changes perception made after feedback.

[krishauser]

Some continued minor changes

[krishauser]

Please put the new attributes in Obstacle, not ObstacleState

[AadarshHegde123]

This file is not authored by Parking team, we copy it from what perception has on their PR branch to ensure consistency. We looked at the file more closely we didn't think having a separate ObstacleState class was necessary.

I have talked with Tianyu and we have agreed to delete ObstacleState, and move anything needed to Obstacle. Please see commit 077ec81

[AadarshHegde123]

We are also up to date with Perception's PR, so this PR merged on its own will include all of perception's updates.

[krishauser]

LGTM

[sonarqubecloud]

Quality Gate failed

Failed conditions
E Reliability Rating on New Code (required ≥ A)

See analysis details on SonarQube Cloud

Catch issues before they fail your Quality Gate with our IDE extension SonarQube for IDE