Mitigation of GMTI Radar False Alarms

Modern GMTI radar systems can detect the movement of unwanted clutter, including wind-blown foliage, waves, and rotating objects. The resulting false alarms create a distraction for radar operators who can miss critical information or improperly allocate resources. Machine Learning (ML) techniques provide a means of classifying detections and discarding these false alarms. In doing so, radar sensitivity is retained, enabling small, slow moving targets (such as dismounts) to be detected near wind-blown tree-lines and other challenging areas. In the Base Phase I effort, Technology Service Corporation (TSC) will focus on developing innovative ML techniques for Army airborne GMTI radars including the Ku-band VADER and X-band LRR systems, and primarily address false alarms caused by wind-blown foliage. TSC will use measured data from these sensors in conjunction with more controlled, simulated data to investigate and evaluate candidate approaches. In the Phase I Option, TSC will refine these ML approaches to mitigate other sources of non-stationary clutter (e.g. waves, rotating antennas and wind turbines) and will extend the algorithms to other sensors including foliage penetration radar. In Phase II, TSC will further mature the most promising approaches and conduct field tests. TSC will also investigate architectures for real-time, on-board processing.