Research on Particle Filter Tracking Method Based on Kalman Filter

Combining real-time of kalman filter or extended kalman filter with robustness of particle filter, proposes a variable particle filter method for target movement with linear motion and non-linear motion. The recommended density function of particle filter is generated by kalman filter when the target do linear motion; The recommended density function of particle filter is generated by extended kalman filter when the target do non-linear motion. The recommended density function makes the latest observation information to integrate into the particle filter in sequential the importance of sampling, and increase the reintegration of likelihood function and the prior distribution, and improve the accuracy of particle filter algorithm. The results of simulation show that the method can effectively reduce tracking error without sacrificing real-time and have a well performance in target tracking.