摘要： 为解决目标跟踪中粒子滤波算法的估计精度、粒子退化问题，提出一种改进的粒子滤波算法。在粒子滤波的基础上，利用UKF生成粒子滤波的建议分布，以改善滤波效果，在无味粒子滤波的基础上，融合典型的MCMC抽样算法，减少传统算法未考虑当前量测对状态的估计作用所带来的影响，增加采样粒子多样化。将该算法应用于具有非线性、非高斯特点的目标跟踪问题中，仿真结果表明，与普通的粒子滤波算法相比，其跟踪精度和滤波效果有较大提高。

关键词: 粒子滤波, 目标跟踪, 非线性滤波, 扩展卡尔曼滤波, 无迹卡尔曼滤波, 马尔可夫链-蒙特卡洛

Abstract: As the problems of estimation accuracy and particles degradation exist in the Particle Filtering(PF) algorithm, an improved PF algorithm is proposed. This algorithm which is based on PF uses the Unscented Kalman Filtering(UKF) to generate the proposal distribution so as to improve the filtering effect. It synchronizes the standard Markov Chain Monte Carlo(MCMC) sampling method and the unscented PF, which can reduce the effect that the traditional PF does not consider the current measurement, and makes the particles more diversification. Simulation results demonstrate that the algorithm has more significant advantages in tracking accuracy and filtering effect than other traditional PF algorithms.

Key words: Particle Filtering(PF), target tracking, nonlinear filtering, Extended Kalman Filtering(EKF), Unscented Kalman filtering(UKF), Markov chain Monte Carlo(MCMC)

中图分类号: 

• TP391