Parameterize odom and body frame IDs

[haasithp]

Summary

This PR replaces the hardcoded TF frame names (camera_init and body) in laserMapping.cpp with ROS 2 parameters. This solves a major pain point by allowing the package to integrate seamlessly into standard robot TF trees (like setting them to odom and base_link) without modifying the source code.

Changes Made

1. Dynamic Frame Variables: Added odom_frame and body_frame variables to laserMapping.cpp to handle TF broadcasts and PointCloud frame IDs dynamically instead of relying on string literals.
2. ROS 2 Parameters: Exposed common.odom_frame and common.body_frame via the parameter server.
3. Reference Config Updated: Added the new parameters to the reference avia.yaml config template so users know how to configure them without guessing.

Backward Compatibility

Fully preserved.
We used get_parameter_or to ensure the defaults perfectly match the old hardcoded values (camera_init and body). Existing users and datasets will experience no changed behavior, and the avia.yaml template uses the exact same frames as before.