A distributed UKF method based on improved joint probability data association
Ning Li, Zhang Zi, Di An · 2017
In this paper, a distributed Unscented Kalman Filter method (UKF) based on improved joint probabilistic data association (IJPDA) is proposed to solve the problem of nonlinear state estimation on multi-sensor multi-target tracking. Firstly, IJPDA method is introduced which determines the source of the multi-sensor data to simplify calculation complexity. Then the distributed UKF algorithm is presented accordingly based on the proposed IJPDA (UKF-IJPDA). Simulations are performed on multiple targets tracking and the results show that the proposed distributed UKF-IJPDA method has lower computational complexity while providing higher accuracy compared with the distributed UKF method based on joint probabilistic data association.