[Feature] Implement Camera-LiDAR Multi-Object Tracking Pipeline

[Ulixes-8]

Description

The multi-modal multi-object tracking pipeline fuses 3D LiDAR detections and Camera-LiDAR detections to form the heart of the multi-modal perception pipeline. It also provides a feedback mechanism between it and the multi-object tracking module to enable smooth, continuous trajectory tracking for dynamic objects over time.

Checklist
Write a launch file that successfully launches:

• Occupancy grid map pipeline.
• LiDAR pipeline.
• Camera-LiDAR pipeline.
• Object association merger module.
• Interpolator module.
• Multi-object tracking module.

Acceptance Criteria

As confirmed by testing:

• The ROI cluster fusion module from the camera-LiDAR pipeline outputs its labeled cluster point cloud as DetectedObjectsWithFeatures to the Interpolator module, which transforms them into DetectedObjects.
• Object association merger module takes in DetectedObjects from both the 3D LiDAR object detection pipeline (from the occupancy grid map-based validator module), as well as the DetectedObjects from the interpolator module.
• The object association merger module outputs DetectedObjects to the multi-object tracker module, which publishes TrackedObjects back to the interpolator module in a feedback loop for continuous detection and tracking.

Priority
High (3/3)