Orthogonal acoustic sensor package for human detection in quiet and noisy environments.

Abstract

Seismic and low-frequency acoustic sensors are severely limited by background vibrations and sound in noisy (urban) areas and a human's ability to mask the impulsive signal from his or her footsteps. Typically, noise levels are highest at low frequencies and roll off at higher frequencies. Recent research has shown that the ultrasonic signals from footsteps can be passively detected in urban areas and in buildings and are relatively independent of attempts at masking. Coupling this sensor with ultrasonic Doppler sonar and processing algorithms that tie footsteps to corresponding human motion yields an orthogonal sensor package with high probability of detection and minimal false alarms. Coupling the ultrasonic sensor package with the low-frequency traditional sensors (such as seismic sensors) forms the universal acoustic sensor package that potentially increases the probability of human detection in high-noise urban and low-noise rural environments. Successful signal processing algorithms for the sensor package are developed and tested. [Work supported by the Department of the Army, Army Research Office, Contract No. W911NF-04-1-0190.]