Anomaly Detection and Intelligent Sensor Resource Management

Advanced UAV-based GMTI radars such as AACER and VADER offer the potential to reliably detect the anomalous behavior of both military vehicles and dismounted troops in real-time to cue narrow field-of-view EO/IR sensors and identify enemy targets amongst civilian vehicle and foot traffic. In Phase I, TSC proposes to explore a GMTI surveillance radar concept for adaptively allocating radar resources to efficiently learn specific regions and schedules of human activity. Models of normal human movement and measured human characteristics will be employed to recognize anomalies. TSC plans to exploit contextual information on populated areas and take advantage of cues including stopped vehicles to control radar waveforms and coverage areas. In Phase I, TSC will develop algorithms to operate in a challenging high density and dynamic target environment such as urban / suburban areas. TSC will extend our innovative anomaly detection algorithms, which were originally developed to recognize unusual vehicle behaviors, to find groups of dismounts which present threats. TSC will modify our existing GMTI radar simulation and scenario generator to develop and test these algorithms, and will then conduct computer experiments to evaluate their performance. In Phase II, TSC will develop a radar resource manager and prototype software for field testing using available GMTI sensors.