A novel nonlinear filter based on f-discrepancy rep-points
Dong Liu, Zhiying Yao, Guangbin Liu · 2006
In practice, the statistics of many systems is usually regarded as Gaussian. Filtering these systems is an attractive field. This paper introduced the F-discrepancy rep-points (FDRP) of multivariate distribution in number theory into the nonlinear filter algorithm, and presented a novel approach for filtering nonlinear Gaussian system based on F-discrepancy rep-points set and Kalman filter (FDRP-KF). The advantage of the method is that there is no parameters needing to set and it can improve the precision increasingly. So it is easy and flexible to use in practice. The results of computer simulations also show that the performance of the algorithm is better than UKF, generic PF. The practical application of the approach also demonstrates its good performance.