观测时滞系统预测滤波算法研究

发布时间：2018-06-17 13:06

本文选题：卡尔曼滤波 + 非线性　；参考：《东北电力大学》2015年硕士论文

【摘要】：在实际目标跟踪应用中，往往会存在由传感器特性引起的时滞现象，该时滞现象是导致系统性能及稳定性变差的主要原因。时滞对目标跟踪精度的影响较为明显，常常导致目标跟踪精度变差，而目前生产、生活甚至军事上对目标跟踪精度的需要越来越高，因此，对时滞的研究具有重要的意义和价值。滤波算法是目标跟踪的重要组成部分，对目标跟踪的精度有着重大的影响。卡尔曼滤波是一种成熟有效的滤波算法，但是对于观测时滞系统中，传统的卡尔曼滤波算法无法直接应用，其关键问题在于观测时滞系统模型中观测方程的形式不符合卡尔曼滤波算法的实现形式。为此，本文在借鉴系统状态增广方法的基础上，提出了一种将观测方程转换为无时滞的观测方程的方法。同时根据卡尔曼滤波算法原理，针对变换后的无时滞的系统模型方程，给出了观测时滞卡尔曼滤波算法步骤。提出的观测时滞卡尔曼滤波算法能够用于解决线性观测时滞系统的预测滤波问题，而提出的观测时滞EKF及UKF算法能够用于非线性观测时滞系统。本文将此方法应用到实际非线性观测光电跟踪系统中，进行算法的性能对比，，仿真实验结果证明，将时滞观测方程转换为无时滞的观测方程的方法是可行的，且可大幅度的提高算法精度。
[Abstract]:In practical target tracking applications, there is often a delay phenomenon caused by sensor characteristics, which is the main reason for the deterioration of system performance and stability. The effect of time delay on target tracking accuracy is obvious, which often leads to the deterioration of target tracking accuracy. At present, the need for target tracking accuracy is increasing in production, life and even military affairs. The study of time delay is of great significance and value. Filtering algorithm is an important part of target tracking, which has a significant impact on the accuracy of target tracking. Kalman filtering is a mature and effective filtering algorithm, but the traditional Kalman filtering algorithm can not be directly applied to time-delay observation systems. The key problem lies in the fact that the form of observation equation in the model of time-delay system does not accord with the realization of Kalman filter algorithm. In this paper, based on the method of system state expansion, a method is proposed to transform the observation equation into one without time delay. At the same time, according to the principle of Kalman filter algorithm and the transformed model equation without time delay, the steps of Kalman filter algorithm with observational time delay are given. The proposed Kalman filtering algorithm with observational delay can be used to solve the problem of predictive filtering for linear observational time-delay systems, while the proposed EKF and UKF algorithms can be used for nonlinear observation time-delay systems. In this paper, the method is applied to the practical nonlinear observation photoelectric tracking system, and the performance of the algorithm is compared. The simulation results show that it is feasible to convert the time-delay observation equation into the time-free observation equation. And can greatly improve the accuracy of the algorithm.
【学位授予单位】：东北电力大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TN713

【参考文献】