基于小波和人工智能的舰载捷联系统动基座对准  

Moving base alignment of shipborne strapdown system based on wavelet and AI technology

作　　者：张涛[1] 徐晓苏[1] 

机构地区：[1]东南大学仪器科学与工程学院,南京210096

出　　处：《吉林大学学报（工学版）》2010年第2期549-553,共5页Journal of Jilin University:Engineering and Technology Edition

基　　金：国家自然科学基金项目(60904088);总装备部预研项目(51309060402;51309020503);中国博士后基金项目(20080441012);航空科学基金项目(20080869009);陕西省电子信息系统重点实验室基金项目(200909A)

摘　　要：将自适应神经模糊逻辑推理系统(ANFIS)应用于舰载捷联惯导系统初始对准中,为INS提供连续的位置、速度和航向信息修正量,提高系统精度并且保证鲁棒性。先利用小波做降噪处理,然后设定INS的位置、速度和航向信息为ANFIS的输入参数,经训练后输出相应修正量,样本的期望输出为经小波多分辨率分析后的结果。仿真结果表明,ANFIS方法水平对准误差在1′内,航向误差在2′内,优于同条件下卡尔曼和RBF的仿真结果。试验说明了该方法能保证舰载捷联惯导系统初始对准的有效性和可靠性。This paper proposes a method of initial alignment of shipborne Strapdown Inertial Navigation System（SINS）based Adaptive Network based Fuzzy Inference Systems（ANFIS）.It provides position,velocity and heading angle corrections for Inertial Navigation System（INS）and improves the accuracy and robustness of the whole system.To achieve this objective,first,wavelet is used to denoise the output of the sensors;then the position,velocity and heading angle of INS are taken as the input of ANFIS;after training the corresponding corrections are output.The expected outputs of trained samples are the results of wave multi-resolution analysis.Test results show that using ANFIS the error of horizontal alignment is within 1＇ and that of azimuth alignment is within 2＇,which are superior to the results of Kalman and RBF under the same conditions.The proposed method can provide reliable and effective alignment for shipborne SINS.

关 键 词：计算机应用 捷联惯性导航 初始对准 小波多分辨率分析 

分 类 号：TP302[自动化与计算机技术—计算机系统结构] V224[自动化与计算机技术—计算机科学与技术]