US Patent 10267912 Audio based motion detection in shared spaces using statistical prediction

An endpoint among a plurality of endpoints, synchronizes a clock across the plurality of endpoints. The endpoint generates a received ultrasonic signal by transducing ultrasonic sound received at a microphone from a spatial region. The ultrasonic sound includes an identical ultrasonic signal transmitted from the plurality of endpoints and echoes from the spatial region. The identical ultrasonic signal is generated with respect to the synchronized clock. The endpoint computes an error signal based on removing the identical ultrasonic signals and the echoes from the received ultrasonic signal. The endpoint detects motion in the spatial region based on a change in the error signal over time.