Consistent Kalman Filter for Distributed Target Tracking with Polar Measurements

This paper deals with a distributed target tracking problem for multiple autonomous vehicles sharing the polar measurements obtained from heterogeneous sensors. To enhance the unsatisfactory convergence property of the existing filtering techniques, a linear consistent Kalman filter (CKF) is proposed. The key idea of our approach is to model the relation between the target state and the polar measurements as a linear uncertain measurement equation. This model has a distinctive form because the uncertain coordinate transform matrix is multiplied, and correlated, to the available measurement matrix. The design objective of the proposed CKF is set to achieve the optimal performance in the sense of estimation error mean even when the coordinate transform uncertainty exists. Different from the previous methods, it provides the consistent target state estimates and makes the heterogeneous sensor fusion easy. Through the simulations, the effectiveness of the proposed method is verified.