safety
The Agent Saw the Obstacle and Still Hit It
Two obstacle-aware harnesses raised collision avoidance to 87.5% without robot-specific training.
Summary
Two obstacle-aware harnesses raised collision avoidance to 87.5% without robot-specific training.
Coding agents could mention a forbidden obstacle in their traces and still treat task completion as the only operational priority. SafeHarness splits the fix into route planning and contact execution: it grounds objects as boxes, proposes and verifies waypoint routes, replans infeasible paths, and chooses contact positions that respect the same constraint. The reported system reached 71.9% task success and 87.5% collision avoidance, 6.5 and 27.0 percentage points above the cited prior state of the art. The authors locate the failure in planning priority rather than in perception or prompt wording.
Why it matters
Two obstacle-aware harnesses raised collision avoidance to 87.5% without robot-specific training.
Limits and context
- Coding agents could mention a forbidden obstacle in their traces and still treat task completion as the only operational priority.
Key claims
Two obstacle-aware harnesses raised collision avoidance to 87.5% without robot-specific training.
Qualification: Coding agents could mention a forbidden obstacle in their traces and still treat task completion as the only operational priority.
Evidence: source-2026-09-18-004
Sources
- arXiv preprint 2609.20822arXiv · primary research
Corrections
No corrections have been recorded for this story.