Coordinated sensing and tracking for mobile camera platforms

This work focuses on developing optimal routing and camera control strategies for multiple mobile PTZ camera platforms cooperatively tracking a stochastic target moving on the ground. The vehicles are modeled as planar Dubins vehicles traveling at a fixed speed and height. The target is tracked using an Extended Kalman Consensus Filter that estimates the target location on the ground plane and the associated error covariance. Using dynamic programming, we compute optimal coordinated control policies which minimize the expected fused geolocation error covariance. The output of this process leads to the trajectories of the vehicles and the PTZ settings of the cameras. We show results on both ground and air vehicles tracking a target whose dynamics is modeled by a stochastic difference equation and discuss the implications of the results. The main contribution of the work is in the development of an optimal strategy for simultaneous vehicle navigation and camera parameter selection to maximize the tracking performance of stochastic targets where the risk of losing the target needs to be considered.