Particle swarm optimization for vehicle positioning based on robust cubature Kalman filter

This paper is concerned with the particle swarm optimization (PSO) based adaptive solution of a robust cubature Kalman filter (RCKF) for data fusion in land vehicle positioning. The cubature rule applied in the cubature Kalman filter is employed to solve the nonlinearity in system and measurement models. With the principle of bounded error covariance, a robust filtering solution is derived to provide tolerance to model uncertainties and the nonlinear approximation errors. On this basis, particle swarm optimization is adopted to assist the original RCKF algorithm and improve the estimation performance by adjusting the restraint coefficient, which is designed as a key parameter. Thus, a novel PSO-assisted robust Kalman filtering scheme is proposed for both the adaption and robustness performance. Simulation analysis and discussions are given to illustrate capabilities and advantages of the proposed filtering solution.