基于修正积分卡尔曼粒子滤波的自适应目标跟踪算法

针对当前粒子滤波权值退化问题以及精度与时耗的矛盾,提出了一种新的高精度自适应粒子滤波算法。该算法综合考虑优选建议分布函数和重采样两种并行改进滤波性能的方法:首先,在积分卡尔曼滤波(QKF)的基础上引入修正因子,通过修正的积分卡尔曼滤波(PQKF)产生优选的建议分布函数,较好地克服了粒子退化现象,在提高滤波精度的同时降低了运算量;在重采样阶段,通过引入系统估计和预测提供的新息差值在线自适应调整采样粒子数,较好地保证了粒子采样的高效性和算法的实时性。实验表明,新算法具有高精度、低时耗的优点,是一种高精度自适应粒子滤波算法。

Adaptive object tracking based on pruning quadrature Kalman particle filter

This paper proposed a new adaptive particle filter for the particle degeneration phenomenon and the contradiction between accuracy and consumption. Considering two parallel improving filtering methods, such as optimum proposed distribution function and re-sampling. First, used pruning quadrature Kalman filter produce optimization proposal distribution function, on the basis of quadrature Kalman filter, introduced pruning factors to improve filtering precision and reduce the running time. In resampling stage, introduced system estimation and prediction for the new spreads value online adaptive adjustments sampling particle counts, kept better sampling efficiency and real-time. Theoretical analysis and experiments show that the proposed new particle filter algorithm has higher accuracy and lower computation time than other improved particle filters, which is a new kind of particle filter algorithm for high precision.