Intra process performace enhancements

[szepilot]

related to #11

• [ ] - DDS implementation (fast dds / cyclone dds)?
• [ ] - Memory leaks?
• [ ] - solve: only WARN when a LiDAR is disconnected
• [ ] - Stress-test
  • [ ] - duration
  • [ ] - demand (ground filter, clustering etc)
• [ ] - Pointcloud output dimension
  • collect use cases to test the code for
• [ ] - Boilerplate code for working with composable node.
• [x] - Workspace building in ros2_ws
• [ ] - Pull request ouster_ros
• [ ] - Documentation
• [ ] - Finally: merge intraproc branch to main branch
• [x] - Real time clustering on merged data
• [x] - Decide wether name is necessary here or not: https://github.com/jkk-research/lexus_bringup/blob/95dfcae31818c53d61e708232deb24c374879c61/launch/drivers/os_32_comp_b.launch.py#L55
• [ ] - Param or other solution to fix 20 Hz loop rate at merger https://github.com/jkk-research/lexus_bringup/blob/64dde7e8ecc90ef5f843f237e926bddafaa2a5c4/src/pcl_merger.cpp#L50
• [ ] Self filter (cuboid filter / CropBox filter) https://github.com/jkk-research/lidar_cluster_ros2/blob/cfcdcfe7d002abc98414e580cbb4ac74d37ee150/src/euclidean_grid.cpp#L268C10-L268C17

[szepilot]

Self filter https://github.com/orgs/jkk-research/projects/5/views/1?pane=issue&itemId=53387479