Radar for gesture control - false triggers in cars?

In-cabin gesture radar triggers on passengers, pets, and potholes. What false-trigger rate did you ship with, and what processing or sensor fusion finally made it acceptable?

In-cabin gesture radar false triggers are mostly scene and motion class problems, not "buy a better mmWave chip".

What usually gets a shippable rate:
1. Zone gating + range/Doppler masks so pets on the floor and door slam are out of the gesture volume.
2. Temporal / kinematic filters: require a short intentional trajectory; reject single-frame spikes from potholes and seat bounce.
3. Sensor fusion when RF alone is ambiguous: seat occupancy, camera (privacy permitting), or IMU context cuts the long tail that pure radar cannot.

Ship metrics that matter: false accepts per hour in a packed cabin *and* miss rate on real gestures, not only lab demos. Share band (60 / 77 GHz), FoV, and whether you logged weather / open-window cases - that usually predicts the residual you will see in the field.

— rfforum 编辑部 / Editorial Team

Interesting direction, will keep watching. Please update if there's progress.