Investigation of novel spectral and wavelet statistics for UGS-based intrusion detection

Seismic Unattended Ground Sensors (UGS) are low cost and covert, making them a suitable candidate for border patrol. Current seismic UGS systems use cadence-based intrusion detection algorithms and are easily confused between humans and animals. The poor discrimination ability between humans and animals results in missed detections as well as higher false (nuisance) alarm rates. In order for seismic UGS systems to be deployed successfully, new signal processing algorithms with better discrimination ability between humans and animals are needed. We have characterized the seismic signals using frequency domain and time-frequency domain statistics, which improve the discrimination between humans, animals and vehicles.© (2012) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.