Add HI-SLAM2 source files, a HI2 specific dockerfile, and slight changes to FASTLIVO

[elliotirving]

Adds HI-SLAM2 to project src/

• With ROSbridge listener run_ros_conversions.py, capable of saving FASTLIVO data to synced frames in a preprocessin/offline fashion (timesynced camera, pose, lidar world points)
• test_ros_conversions.py to verify preprocessed data
• Still need to implement online version with your custom 3DGS module

Adds HI-SLAM2 dockerfile

• Base Docker environment has all FASTLIVO dependencies, use act_hi2 inside container terminal to activate HI-SLAM2 conda environment on top of this. NOTE this is because HI2 uses newer Python than ROS Noetic for FASTLIVO.
• Still untested GPU pass through capability, as my local one is overloaded running original HI-SLAM2 module

Slight modification to FASTLIVO

• publish odometry in handleVIO() function to ensure adequate frame time sync