Model noise Kalman filter algorithm for electro-optical tracking

In order to reduce the sensitivity of the filtering performance to the model noise in the target state space model,the method of updating the system noise covariance and measurement noise variance was proposed.By combining it with nonlinear Kalman filter,an adaptive nonlinear Kalman filter was constituted which is suitable for applying in electro-optical target tracking.Meanwhile it was applied in nonlinear measurement electro-optical tracking system,and compared with those of extended Kalman filter and unscented Kalman filter.The Matlab simulation results show that this method can adjust measurement noise covariance and system noise covariance in real time,and can effectively avoid the problem of filter performance degradation caused by the inaccurately statistical properties of the measurement noise and system noise,and the performance by the model noise Kalman filter significantly outperforms those of the extended Kalman filter and the unscented Kalman filter.