EKF在机器人目标跟踪中的应用  

Application of EKF in Robot Target Tracking

作　　者：温彬彬 温鹏智 胡雪花 WEN Bin-bin;WEN Peng-zhi;HU Xue-hua(Hebei Vocational University of Industry and Technology,Shijiazhuang 050000,China;Xingtai Modern Occupation School,Xingtai 054000,China)

机构地区：[1]河北工业职业技术大学,河北石家庄050000 [2]邢台现代职业学校,河北邢台054000

出　　处：《电脑知识与技术》2021年第36期116-118,共3页Computer Knowledge and Technology

基　　金：河北工业职业技术大学院级自然科学课题(课题编号:zky2020002);2021年度河北省高等学校科学技术研究项目(项目编号:ZC2021017)。

摘　　要：扩展卡尔曼滤波在机器人目标跟踪中有非常重要的应用,该文重点介绍了扩展卡尔曼滤波的工作流程以及如何应用在机器人目标跟踪中。但EKF算法应用在非线性系统中会产生二次项以上的截断误差,累计会影响机器人的定位精度。该文主要从拓展卡尔曼滤波的收敛性上进行分析,利用双估计用观测残差对过程噪声进行补偿,利用新息和观测值估测观测噪声矩阵,通过多次预测、更新迭代从而达到对机器人更精准的定位。Extended Kalman filter has a very important application in robot target tracking. This paper focuses on the workflow of extended Kalman filter and how to apply it in robot target tracking. However, when EKF algorithm is applied in nonlinear system, it will produce truncation errors above quadratic term, which will affect the positioning accuracy of the robot. This paper mainly analyzes the convergence of extended Kalman filter, uses double estimation to compensate the process noise with observation residuals, uses innovation and observation values to estimate the observation noise matrix, and achieves more accurate positioning of the robot through multiple prediction and update iterations.

关 键 词：扩展卡尔曼滤波 目标跟踪 噪声补偿 观测噪声 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]