基于 Box 粒子滤波的参数自适应机动目标跟踪

摘要

针对机动目标跟踪加速度的不确定性，引入一种新的参数自适应算法，采用粒子滤波及高斯核密度估计技术，估计目标机动参数，实现对任意机动目标的跟踪。在此基础上，考虑到粒子滤波计算代价较高的问题，进一步引入区间分析技术，采用 Box 粒子代替传统的粒子，以提高算法的计算效率。实验结果表明，提出的算法能够有效地跟踪任意机动目标，且运算时间明显低于传统的参数自适应算法。%A new kind of adaptive parameter estimation algorithm is introduced for the acceleration uncertainty of ma-neuvering target tracking.The particle filter and Gaussian kernel density estimation are used to estimate target maneu-vering parameter accurately.In order to decrease the computational complexity,the interval analysis technology is fur-ther introduced.Box particles are used to replace traditional particles,which can improve the computational efficien-cy.The simulation results show that the proposed algorithm can effectively track arbitrary maneuvering target,and the computation time is significantly less than that of the traditional adaptive parameter estimation algorithm.