Behaviour Recognition of Ground Vehicle for Airborne Monitoring by UAV Swarm

This paper focuses on development of behaviour recognition technique for airborne monitoring of ground traffic to detect hidden threats. To enhance tracking accuracy, sensor fusion and smoothing are applied with Kalman filter. To tackle behaviour recognition, trajectory approximation and classification methodology is proposed using differential geometric quantities and string matching. Path planning strategy is also touched for tracking a moving ground vehicle. Simulation on a ground vehicle is done to verify the feasibility of the proposed algorithms.