Summary
Target rates are:
- LiDAR: 20 Hz
- VectorNav: 100 Hz
Current status: as noted in Issue #1, both are running at 20 Hz right now.
Once VectorNav is restored to 100 Hz, we need to decide the correct recording architecture.
At present, both streams are logged into a single rosbag file.
Problem to solve
When LiDAR and VectorNav run at different frequencies (20 vs 100), we need to decide whether to:
- Keep recording both into a single rosbag, or
- Record into two separate rosbags in the same session folder, or
- Support both modes via a runtime flag/option in
driver_node.
Context
- Single-file logging is simpler for replay and session management.
- Split-file logging may improve post-processing flexibility, and per-sensor workflows.
- We currently do not have enough usage data to lock this decision.
Proposed direction
- Evaluate both approaches with representative workloads.
- Add a configurable run option (e.g.
--bag-mode single|split) so behavior can be selected per use case.
- Keep current behavior as default unless testing shows a clear reason to switch.
Acceptance criteria
- Decision documented with pros/cons and rationale.
- Recording strategy implemented (single, split, or configurable mode).
- If configurable mode is chosen:
- CLI option is added and documented.
- Output folder structure is clearly defined.
- Verified recording works correctly with LiDAR at 20 Hz and VectorNav at 100 Hz.
Summary
Target rates are:
Current status: as noted in Issue #1, both are running at 20 Hz right now.
Once VectorNav is restored to 100 Hz, we need to decide the correct recording architecture.
At present, both streams are logged into a single rosbag file.
Problem to solve
When LiDAR and VectorNav run at different frequencies (20 vs 100), we need to decide whether to:
driver_node.Context
Proposed direction
--bag-mode single|split) so behavior can be selected per use case.Acceptance criteria