Collaborative Human-Animal Discrimination for Unattended Ground Sensors

Effective human-animal discrimination happens collectively at all levels of an UGS system. The ability to classify a target from ootsteps alone as human versus non-human given a single modality and sensor is challenging, thus a collaborative network effort increases the system Probability of Detection (Pd) and minimizes Probability of False Alarm (Pfa). IWT proposes leveraging inter-UGS communication to both increase local classification probabilities and provide a system level indication of target intent. Humans rossing the border, cross with a purpose. This trait differs from animals loitering or stationary vegetation swaying in the wind. Correlation of information between UGS nodes provides strong target behavioral information on location, rate, and direction of travel as well as enhancing the alarm statistics of individual UGS units.