Sensor Redundancy Rig

Toggle lidar, radar, and camera coverage while the autonomy stack encounters fog, construction detours, occlusions, and emergency vehicles. Each scenario mirrors event tallies in the DMV disengagement narratives: see how composite confidence collapses when redundancy is stripped away.

Confidence is computed by applying scenario-specific reliability weights to each active sensor, then combining them with Bayesian fusion. Numbers cite CA DMV 2023 narratives (Waymo, Cruise, Zoox, Aurora).

Scenario

Lidar High-resolution depth map Radar Long-range velocity / adverse weather anchor Cameras Semantic understanding & color cues

Composite detection confidence

—%

Enable sensors to compute

Expected disengagement rate

— per 10k mi

Scenario-specific baseline from DMV reports