基于粒子滤波的非线性目标跟踪算法研究

针对非线性非高斯的目标跟踪，传统的卡尔曼滤波和扩展卡尔曼滤波等算法将会出现滤波精度下降甚至发散的现象，提出了采用粒子滤波算法来解决非线性滤波问题；粒子滤波方法作为一种基于贝叶斯估计的非线性滤波算法，在处理非高斯非线性时变系统的参数估计和状态滤波问题方面有独到的优势，但是存在运算量大和实时性差的问题，因此提出了基于EKF的扩展粒子滤波；仿真结果表明：在强非线性非高斯环境下，PF算法的跟踪性能优于EKF算法，基于EKF的扩展粒子滤波能够取得较好的跟踪精度，并且能够有效的减少粒子滤波的运算量。

Nonlinear Target Tracking Algorithm Based on Particle Filters

Particle Filter is presented to solve the nonlinear filter and non-Gaussian problem, while the al- gorithms of Kalman Filter and Extended Kalman Filter within the Gaussian background leads to the filter precision decrease and divergence phenomenon. As a nonlinear filter algorithm based on Bayesian estima- tion, particle filter has original advantage at treating the parameter estimation and state filtering aspects of nonlinear non-Gaussian time-varying systems, but it takes a lot of time due to larger number of particles. Thereby Extended Kalman Particle Filter is presented to solve the lower the real-time performance resulting from high computational complexity. The simulation results show that the PF approach outperforms the EKF algorithm under strong nonlinear and non-Gaussian environment, and EKPF gives better performance than EKF in solving high computation complexity.