Application Of Wavelet Particle Filter In The Single Manuevering Target Tracking

The key factor in Target tracking is the tracking model and the filter algorithm. In the Single maneuvering target model,there are CV model CA model Singer model Current statistical model Coordinate Turn model and so on.The main filter agorithm is standard Kalman filter algorithm Extended Kalman algorithm Unsented Kalman filter and In recent year, particle fliter algorithm besed on byes theory was put forward.the filter agorithm can be devided into linear agorithm and nonlinear agorithm.In linear filter agorithm,the maneuvering target model is the key factor.Especially,the algorithm like Kalman filter. Only the maneuvering target model match the target movement,the agorithm can arrive convergence as soon as possible and rise the filter precision.In nonlinear filter agorithm,this passage analyse the advantage of nonlinear agorithm in theory.In chapter 3 of this paper,the main emphasis is particle filter agorithm in theory.The particle filter evolve from byes estimate.In practice filter the main problem is the particle degenerate.Usually the resampling the particle can elevate the performance of particle filter.Therefore,some improved particle filter come out.For example APF RPF EKF-PF,and so on. According to the thoery of particle filter and Single maneuvering target tracking, In this chapter put forward the diagram of particle filter in the Single maneuvering target tracking.At the end of this chapter,comparing the efficence of the particle filter and the EKF agorithm .According to the re-entry phase of the ballistic missile model, make the simulation to test the particle filter and the improved particle filter. In the chapter 4,introduceing the thoery of wavelet,analying the superiority of wavelet analysis and the ability of its denosing .When the particle filter is used in the target tracking,There is an error between the particle sampled and the postior posibability.This passage put forward using the wavelet to denosing this error.Using wavelet to denoise the particle weight,it decrease the error between the particle and the real postior probability.The simulation result shows wavelet particle filter has the lower RMSE than standard particle filter.