Study On Performance Analysis Of Nonlinear Filtering Algorithms For Target Tracking

Target tracking is an important theoretical and practical problem.Actually,target tracking problem is a tracking and filtering problem of target state,that is to estimate target state accurately according to the measurement data achieved by sensors.Nonlinear filtering is an important aspect of target tracking.In our real world,all of the control systems are nonlinear.Linear is an approximate description to nonlinear in some degree.Along with the development of technology,the require for estimation accuracy and tracking speed is more and more high for target.Accordingly,looking for more credibility and efficient filtering algorithm is still an important problem.Aim at this problem,this dissertation study the nonlinear filtering method in target tracking deeply.Firstly,the research background and meaning are introduced;and then,the basic principle of target tracking and the situation of study are described simply.Furthermore,nonlinear filtering algorithm and its application are summarized in detail.At last,research fruits of this dissertation are given.The main contributions are as follows: 1.Estimation Research Based on Kalman Filter AlgorithmThe Kalman filter algorithm is mainly studied.The Kalman filter based on S-function is designed and simulated.The adaptive noise estimation based on Kalman filter is studied.The joint Kalman filtering algorithm is described in detail,and the simulation results show that the joint Kalman filtering algorithm can be applied to the system in which its target state and the control input are unknown,and can achieve convergence in a relatively short time.2.Application of improved particle filter algorithm in target trackingThe algorithm flow of CPF filtering algorithm,EPF filtering algorithm and UPF algorithms are introduced respectively.The advantages and disadvantages of each algorithm are analyzed and summarized.It is found that UPF algorithm has the best filtering effect among the three filtering algorithms.Then the improved MUPF filtering algorithm is studied.The simulation results show that the improved MUPF algorithm has better filtering performance than UPF algorithm.3.Performance Analysis of Improved Square Root Cubature Kalman FilterThe cubature Kalman filter(CKF)algorithm and its improved square root cubature Kalman filter(SCKF)algorithm are elaborated in detail.A new nonlinear-linear square root cubature Kalman filter(NL-SCKF)algorithm is proposed.The simulation results show that the improved SCKF and NL-SCKF algorithm have higher filtering accuracy than CKF algorithm,and NL-SCKF algorithm has shorter operation than SCKF algorithm.