Particle Filter And Its Improved Algorithm In Target Tracking

The core issue of target tracking is the state estimation in the complex interference environment and the condition of intensive target, including course, azimuth, component motion parameters such as location, speed, acceleration and so on. The target implementation of accurate location and tracking in the modern military system is the pressing needs of the modernization of national defense construction and national security in order to ensure better finish the task. The research work of this article is studyding Filtering algorithm of target tracking system for theoretical research and experimental simulation in this context.In a linear system,mainly using least squares or kalman filter method to solve the problem of dynamic system of target tracking; In nonlinear systems, the research hotspot in recent years focused on improved algorithm based on kalman filter and particle filter (PF) and its improved algorithm.PF algorithm gradually caused the attention of the researchers because its merits, such as not effected by the nonlinear degree of the the state equation and measurement equation and unlimited the noise.As the modern military and civilian fields for target tracking accuracy demand continues to improve, so the main contents of this paper is to improve filtering accuracy, and made two improved filtering algorithm:Considering the problem of poor tracking accuracy and particle degradation in the traditional Particle Filter (PF) algorithm for high-performance requirements of the target tracking, this article discussed a new improved nonlinear filtering algorithm based on Unscented Kalman Particle Filter (UPF). The improvement of the algorithm use the latest observation information into the Unscented Kalman Filter(UKF) to generate the importance density function in UKF links, adding the scaling factor can dynamic mediation parameter to solve the problem of sampling the local effects; using the system resample in the link of PF resampling, According to the effective particle volume set weight threshold to determine the need for resampling, finally make a new algorithm called System Proportion Symmetry Unscented Particle Filter(SPSUPF). Simulation results show that SPSUPF can effectively solve the problem of degradation of particles in the particle filter and improve filtering accuracy when the time-consuming considerable.In order to slove the problem of the proposal distribution function selecting of the traditional particle filter algorithm and the particle degradation phenomenon, an improvement particle filter algorithm based on Markov Monte Carlo ideas was proposed. Firstly, the algorithm applies the unscented Kalman filter that makes use of the proportional symmetry sampling methods emerge the Sigma points to generate the recommendations distribution of particle filter; Secondly, applying the likelihood distribution adaptive strategy in weights selection procedure; Finally, applying the system resampling methods and joining Markov chain Monte Carlo procedure to maintain particle diversity. Simulation results show that WAUPF-MCMC can effectively inhibit the particle deprivation and has the merits of tracking accuracy than the traditional non-linear filters.