feat(prediction): implement module skeleton with multi-modal architec…

[RyanL2]

📑 Description

This PR implements the foundational skeleton for the prediction module within the world_modeling stack. It establishes the core node architecture and defines the necessary interfaces for multi-modal trajectory prediction (Vehicle, Pedestrian, Cyclist).

This structure is needed to:

Enable Parallel Development: Decouples the work for team members (John, Girish, Aruhant) so they can implement specific motion models simultaneously without merge conflicts.

Standardize Interfaces: Enforces a consistent TrajectoryHypothesis output format across all object types.

Provide a Runnable Base: Delivers a compiling and runnable ROS 2 node (using placeholder logic and raw pointers for safe memory management) that can be tested immediately.

📹 (Optional) Video Demo of Changes

✅ Checklist

• My code builds and runs locally without warnings
• I added/updated tests if needed
• I updated documentation / comments
• I listed any breaking changes in the “Notes” section

🔗 Related Issues / PRs

📝 Notes for reviewers

[Copilot]

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.

[Edwardius]

nice work! things are skeleton state atm so im not going to review too heavily. careful about ai generated slop, usually it tends to make unnessesary comments (ie. comments of things being removed, if it doesn't exist, why comment about it?)

As for architecture, it seems sane on paper.

[Edwardius]

use yaml launchfile format instead please

[RyanL2]

Do you mean converting the Python launch file to XML format?