组合体航天器角速度信号深度学习降噪方法  

ANGULAR RATE SIGNALS DENOISING METHOD OF THE COMBINED SPACECRAFT BASED ON DEEP LEARNING

作　　者：初未萌 邬树楠[1,2] 刘宇飞[3] 吴志刚[1,2] Chu Weimeng;Wu Shunan;Liu Yufei;Wu Zhigang(School of Aeronautics and Astronautics,Dalian University of Technology,Dalian 116024,China;State Key Laboratory of Structural Analysis for Industrial Equipment,Dalian University of Technology,Dalian 116024,China;China Academy of Space Technology,Qian Xuesen Laboratory of Space Technology,Beijing 100094,China)

机构地区：[1]大连理工大学航空航天学院,大连116024 [2]大连理工大学工业装备结构分析国家重点实验室,大连116024 [3]中国空间技术研究院钱学森空间技术实验室,北京100094

出　　处：《动力学与控制学报》2019年第6期584-592,共9页Journal of Dynamics and Control

基　　金：国家自然科学基金资助项目(11502040);大连市高层次人才创新支持计划项目(2017RQ001)~~

摘　　要：服务航天器与非合作目标构成组合体航天器后,实际测量的角速度信号往往掺杂了较大噪声,这会严重影响后续控制的精度,但由于非合作目标惯性张量未知,导致无法根据模型信息对其有效降噪.针对上述问题,本文基于深度学习方法提出一种无需预先构建模型、完全由数据驱动的组合体航天器角速度信号降噪方法.首先给出采用深度学习方法对角速度进行降噪所需训练数据的生成过程;然后构建组合体航天器角速度信号降噪的深度网络模型,并提出该模型的优化方法和参数初始化方法;最后用测试数据对本文方法和小波降噪方法的降噪效果进行了比较,结果表明本文方法与小波降噪方法相比具有更好的降噪性能.For the combined spacecraft consisting of the servicing spacecraft and non-cooperative target,the measured angular rate signals of the combined spacecraft are often affected by large noises,which will influence the sub-sequent control accuracy seriously.Due to the unknown inertia tensor of the non-cooperative target,the signal cannot be effectively denoised using model information.To solve the above-mentioned problem,an angular rate signals denoising method of the combined spacecraft based on the deep learning method is proposed in this paper,which do not require the model information in advance,and is absolutely data-driven.Based on the deep learning,the generative process of the training data which is needed for angular rate denoising is firstly presented,and then a deep network model for angular rate signals denoising of the combined spacecraft is constructed,including the methods of optimization and parameter initialization of the model.Finally,denoising effects are compared between the proposed method and the wavelet denoising method using test data.The result shows that the proposed deep learning method has a better denoising effect than the wavelet method.

关 键 词：深度学习 组合体航天器 非合作目标 信号降噪 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]