Abstract: Aiming at the problem of targe tracking in Bayesian filtering process, some typical Bayesian filtering methods such as EKF, UKF, PF and UPF are proposed. On basis of the frameworks set up by these methods, the performance of these methods are tested and compared, and the characteristics of them are discussed in non-linear environment. Simulation experimental results show that, in non-linear and non-Gaussian environment, the performance of UPF is the best.

Key words: target tracking, Bayesian filtering, non-linear filtering method

摘要： 针对贝叶斯滤波过程中存在的目标跟踪问题，提出几种典型的贝叶斯滤波方法，如EKF, UKF, PF和UPF等，基于这些方法所构建的框架，对它们进行性能测试和比较，并在非线性环境下，讨论这些方法的特点，仿真实验结果表明，在非线性非高斯环境下，UPF方法的性能是最优的。

关键词: 目标跟踪, 贝叶斯滤波, 非线性滤波方法

CLC Number: 

• TP391