Seismic Source Discrimination

Surveillance of remote border areas is a significant problem. Detection of walking or running subjects is currently performed using
multi-modal unattended ground sensors that measure seismic, acoustic, passive infrared (PIR), and magnetic activity. Current systems rely primarily on the seismic and PIR sensors as they provide the greatest detection range. However, these sensors are prone to false
alarms as they can not effectively distinguish between human and animal activity. These false alarms are particularly problematic in remote areas due to the manpower required to investigate each alarm and the abundance of wildlife. During the course of this project, Creare will continue to develop and refine our novel feature-based intruder detection and discrimination algorithms. These algorithms, designed for use in unattended ground sensors, have already shown promising results
when used to discriminate between human and animal data.