文档介绍：第56卷第4期 西安交通大学学报 Vol. 56 No. 4
2022年4月 JOURNALd is used to
extend the rule into multivariate quadrature rule by for the numerical approximation of multi­
dimensional integrals； Finally, the multivariate quadrature rule is introduced into the framework
of nonlinear Gaussian recursive filtering, and the SGKQF is obtained by implementing the square­
root form of the covariance in the filtering process to enhance the stability. The proposed
algorithm is applied to the bearings-only target tracking system to test its performance. The
simulation result shows that，compared with traditional GHQF algorithm, SGKQF can achieve
higher filter estimation accuracy under the condition of selecting appropriate Gaussian kernel
bandwidth, with the improvement of estimation accuracy in position and velocity by 8・ 7% and
11・ 8 % respectively・
收稿日期：2021-08-05 o 作者简介：汪家宝(1996-),男，硕士生；吴昊(通信作者)，男，讲师。 基金项目：国家自然科学
基金资助项目(62073337)；陕西省自然科学基础研究计划资助项目(2020JQ-479) 。
网络出版时间：2021-11-16 网络出版地址：http： //kns. cnki. net/kcms/detail/61. 1069. T. 20211115. 2202. 006. html158 西安交通大学学报 第56卷
Keywords: nonlinear system； target tracking； Gauss-Hermite quadrature rule；