基于随机矩阵建模的Student’s t逆Wishart滤波器

Student’s t inverse Wishart filter based on random matrix modeling

中文关键词  扩展目标跟踪  随机矩阵模型  逆Wishart分布  异常噪声

英文关键词  extended target tracking  random matrix model  inverse Wishart distribution  abnormal noise

中文摘要

      针对复杂的异常噪声条件下扩展目标跟踪问题, 本文基于随机矩阵模型(RMM)提出了一种Student’s t逆 Wishart滤波算法. 首先, 运用Student’s t分布对异常的过程噪声和量测噪声进行建模, 运用逆Wishart分布对目标扩展 状态进行建模, 以更加合理的数学模型表征异常噪声条件下基于稀疏量测的目标基本轮廓特征. 进一步的, 本文详 细推导了能稳健估计目标椭圆形状特征的Student’s t逆Wishart滤波算法, 能在形状和方向动态演变过程中有效估计 扩展目标的多重特征. 最后, 通过构造异常噪声条件下椭圆扩展目标跟踪的仿真实验验证了所提算法的有效性.

英文摘要

      For the issue of the extended target tracking in the complex abnormal noise conditions, a Student’s t inverse Wishart filtering algorithm using random matrix model (RMM) is proposed in this article. First, the Student’s t distribution is used to model the abnormal process noise and the measurement noise, and the inverse Wishart distribution is used to model the extended state of target. Thus, a more reasonable mathematical model is used to represent the basic contour features of target based on the sparse measurement with the abnormal noise. Furthermore, this paper deduces a Student’s inverse Wishart filtering algorithm which can robustly estimate the ellipse shape of target, and can effectively estimate the multiple characteristics of the extended target with the dynamic evolution of the shape and the direction. Finally, the simulation experiment about elliptical extended target tracking with the abnormal noise verifies the effectiveness of the proposed algorithm.