基于多伯努利滤波的厚尾噪声条件下多扩展目标跟踪

针对厚尾噪声条件下不规则星凸形多扩展目标跟踪(Multiple extended target tracking, METT)问题, 提出一种基于多伯努利滤波的厚尾噪声条件下多扩展目标跟踪方法. 首先, 采用学生t分布对厚尾过程噪声和量测噪声进行建模, 并基于有限集统计(Finite set statistics, FISST)理论利用随机超曲面模型(Random hypersurface model, RHM)建立不规则星凸形多扩展目标的跟踪滤波模型. 然后, 利用学生t混合(Student＇s t mixture, STM)模型来表征多伯努利密度, 提出学生t混合多扩展目标多伯努利(Student＇s t mixture multiple extended target multi-Bernoulli filter, STM-MET-CBMeMBer)滤波算法, 并进一步基于鲁棒学生t容积滤波算法提出了非线性鲁棒学生t混合星凸形多扩展目标多伯努利滤波算法. 最后, 通过构造厚尾噪声条件下星凸形多扩展目标和多群目标的跟踪仿真实验验证了所提方法的有效性.

Multiple Extended Target Tracking in the Presence of Heavy-Tailed Noise Using Multi-Bernoulli Filtering Method

Aiming at the problem of the irregular star-convex multiple extended target tracking (METT) with heavy-tailed noise, a multiple extended target tracking in the presence of heavy-tailed noise using multi-Bernoulli filtering method is proposed in this article. First, the student＇s t distribution is used to model the heavy-tailed process noise and measurement noise. The irregular star-convex multiple extended target filtering problem is formulated based on the finite set statistics (FISST) theory and the random hypersurface model (RHM). Then, the multi-Bernoulli density is approximated by the student＇s t mixture (STM) and a student＇s t mixture multiple extended target multi-Bernoulli filter (STM-MET-CBMeMBer) is proposed correspondingly. Furthermore, this article proposes a nonlinear robust student＇s t mixture star-convex multiple extended target multi-Bernoulli filter based on the robust student＇s t based cubature filtering method. Finally, simulation experiments on star-convex multiple extended target tracking and multiple group target tracking with the heavy-tail noise verify effectiveness of the proposed method.