Novel Interacting Multiple Model Filter for Uncertain Target Tracking Systems Based on Weighted Kullback–Leibler Divergence
Bowen Hou,Jiongqi Wang,Zhangming He,Yongrui Qin,Haiyin Zhou,Dayi Wang,Dong Li
DOI: https://doi.org/10.1016/j.jfranklin.2020.09.012
IF: 4.246
2020-01-01
Journal of the Franklin Institute
Abstract:Interacting multiple model (IMM) filter is a classical method to track targets in hybrid situations. However, it can exhibit divergence when the models are correlated or the system suffers from uncertainties. The generalized covariance intersection method based on the weighted Kullback–Leibler (K–L) divergence can solve the divergence problem of correlated estimates. A novel interacting multiple model (NIMM) filter is presented that combines two different algorithms, the adaptive fading Kalman filter and the maximum correntropy Kalman filter, based on the model interacting with the weighted K-L divergence to address the uncertainty problems of the system. The NIMM filter algorithm is designed and the stability and accuracy are analyzed. The simulation results demonstrate that the proposed filter can effectively improve the accuracy under different uncertainty conditions for classical examples and ballistic trajectory tracking scenarios.