基于交互多模型的机动目标跟踪算法研究

发布时间：2018-02-20 21:23

本文关键词： 机动目标跟踪非线性滤波交互多模型协方差矩阵交互多模型　出处：《大连海事大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着现代航运业的快速发展,水上交通形势日趋严峻,船舶安全航行问题日益突出,从而对船舶跟踪性能提出了更高的要求,特别是机动目标的稳定跟踪更为关键。为改善目标跟踪可靠性,提高机动目标跟踪精度。本文在传统交互多模型算法(IMM)的基础上,展开对机动目标跟踪算法的研究。首先,本文概括了机动目标跟踪的基本原理,介绍了常用的目标运动模型。然后在卡尔曼滤波的基础上重点研究了扩展卡尔曼滤波(EKF)和无迹卡尔曼滤波(UKF)这两种主要的非线性滤波算法。并设计出目标机动运动场景,仿真对比分析了 EKF算法和UKF算法的机动目标跟踪性能,为后文的改进算法提供了理论依据。其次,为了适应目标的机动变化,在IMM算法原理及优缺点进行分析的基础上,将IMM算法分别与EKF和UKF这两种非线性滤波算法结合,研究并设计了基于扩展卡尔曼滤波的交互多模型算法(IMM-EKF)和基于无迹卡尔曼的交互多模型算法(IMM-UKF)。针对目标匀速和转向运动的运动场景,分别对EKF、UKF、IMM-EKF和IMM-UKF四种目标跟踪算法进行仿真分析。与传统算法相比,基于非线性滤波的IMM算法提高了机动目标的跟踪精度。最后,为提高模型概率的准确率,研究了已有的基于模型概率修正的交互多模型算法(SIMMP),分析了该算法模型概率的修正过程,并根据基于将当前协方差信息引入模型概率的思想,本文提出了一种基于协方差矩阵修正概率的交互多模型算法(MIMMP),并对此算法与IMM算法和SIMMP算法进行仿真对比分析,证明了本文算法的有效性和优越性。
[Abstract]:With the rapid development of modern shipping industry, the situation of water transportation is becoming more and more serious, and the problem of ship safe navigation is becoming more and more prominent, which puts forward higher requirements for the performance of ship tracking. In order to improve the reliability of target tracking and improve the precision of maneuvering target tracking, based on the traditional interactive multi-model algorithm (IMM), the maneuvering target tracking algorithm is studied in this paper. In this paper, the basic principle of maneuvering target tracking is summarized. Based on the Kalman filter, two main nonlinear filtering algorithms, the extended Kalman filter (EKF) and the unscented Kalman filter (UKF), are introduced, and the maneuvering moving scene of the target is designed. The simulation results show that the maneuvering target tracking performance of EKF algorithm and UKF algorithm is compared and analyzed, which provides a theoretical basis for the improved algorithm. Secondly, in order to adapt to the maneuvering change of target, the principle of IMM algorithm and its advantages and disadvantages are analyzed. The IMM algorithm is combined with two nonlinear filtering algorithms, EKF and UKF, respectively. This paper studies and designs an interactive multi-model algorithm based on extended Kalman filter (IMM-EKF) and an interactive multi-model algorithm based on unscented Kalman filter (IMM-UKF). Four target tracking algorithms are simulated and analyzed. Compared with the traditional algorithm, IMM algorithm based on nonlinear filtering improves the tracking accuracy of maneuvering target. Finally, in order to improve the accuracy of model probability, the tracking accuracy of maneuvering target is improved. In this paper, the existing interactive multi-model algorithms based on model probability correction are studied, and the process of model probability correction is analyzed. Based on the idea of introducing the current covariance information into the model probability, In this paper, an interactive multi-model algorithm based on the modified probability of covariance matrix is proposed. The simulation results of this algorithm are compared with that of IMM algorithm and SIMMP algorithm. The results show that the proposed algorithm is effective and superior.
【学位授予单位】：大连海事大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN953

【参考文献】