Use when an LLM, VLM, learned policy, or any AI model issues commands to a physical robot (mobile base, arm, drone, rover). Provides the complete guardrail architecture: validated command schemas so AI never touches raw motors, velocity/accel/jerk clamps below the AI layer, geofence enforcement, hum
Use when an LLM, VLM, learned policy, or any AI model issues commands to a physical robot (mobile base, arm, drone, rover). Provides the complete guardrail architecture: validated command schemas so AI never touches raw motors, velocity/accel/jerk clamps below the AI layer, geofence enforcement, human override priority chains, command rate limiting, sim pre-validation of AI plans, runaway detectio