Affiliations: School of Electrical and Information Engineering, Zhengzhou University of Light Industry, Zhengzhou, China
Correspondence:
[*]
Corresponding author. Yuzhao Jiao, School of Electrical and Information Engineering, Zhengzhou University of Light Industry, Zhengzhou, China. E-mail: jiaoyuzhao@126.com.
Abstract: For multi-sensor target tracking system, the accurate state estimation is obtained under the condition that the system model is unbiased and the noise disturbance satisfies the characteristics of independent Gaussian white noise. However, in engineering practice, it is almost impossible to get the accurate system model and also the system noise is non-independent Gaussian white noise. So the traditional state estimation methods are not suitable for uncertainty system with non Gaussian white noise. In this paper, the Kalman Filter-Support Vector Machine (KF-SVM) based data fusion algorithm is proposed for system with model uncertainty and correlated noise. Firstly, the state pre-estimates are calculated by the proposed improved Kalman Filter for single sensor system. Then, the state estimation is obtained via proposed KF-SVM data fusion algorithm for multi-sensor system. Finally, compared with the traditional algorithms, the simulation results show that the proposed fusion algorithm based on KF-SVM does not need to calculate the sensor cross-covariance matrix and has better estimation accuracy.
Keywords: Support vector machine, kalman filter, data fusion, system uncertainty, cross-correlated noise
