Safety layer for VLA models and humanoid robots. Sub-ms physics guard, ISO-aligned, zero neural nets.
-
Updated
Aug 15, 2026 - Python
Safety layer for VLA models and humanoid robots. Sub-ms physics guard, ISO-aligned, zero neural nets.
Deterministic physics safety for embodied AI — beyond statistical safety guards.
ISO 10218 and ISO/TS 15066 safety standards explained, with deterministic physics safety validation.
Deterministic Newtonian safety engine for robotics — force, speed, contact area, all computed.
Dynamic contact area modeling for robot collision safety — pressure, not just force.
Physics-based collision detection with dynamic contact area, impulse boundary, and 7-level risk grading.
Whole-body safety for humanoid robots — balance, fall detection, and contact physics.
500+ labeled VLA safety scenarios across 7 risk levels — benchmark dataset for safety evaluation.
To associate your repository with the dynamic-contact-area topic, visit your repo's landing page and select "manage topics."