\r窦立谦¹，杜苗苗¹，张秀云¹，王跃萍\r²\r
\r
AuthorsHTML:\r窦立谦¹，杜苗苗¹，张秀云¹，王跃萍\r²\r
\r
AuthorsListE:\rDou Liqian¹，Du Miaomiao¹，Zhang Xiuyun¹，Wang Yueping\r²\r
\r
AuthorsHTMLE:\rDou Liqian¹，Du Miaomiao¹，Zhang Xiuyun¹，Wang Yueping\r²\r
\r
Unit:\r1. 天津大学电气自动化与信息工程学院，天津 300072；
2. 飞行控制航空科技重点实验室(航空工业自控所)，西安 710065\r
\r
Unit_EngLish:\r1. School of Electrical and Information Engineering，Tianjin University，Tianjin 300072，China；
2. Aviation Key laboratory of Science and Technology on Aircraft Control，FACRI，Xi’an 710065，China\r
\r
Abstract_Chinese:\r可重复使用运载器(RLV)再入返回段的气动参数表现为不确定和快时变的特点，导致RLV 气动特性具有强耦合性和非线性，使气动模型难以设计和控制，降低了飞行器的稳定性．在系统噪声存在的前提下，针对RLV 的动力学模型，提出一种由最大期望(EM)算法和扩展卡尔曼滤波(EKF)算法相结合的RLV 再入段气动参数辨识方法，以飞行高度和攻角为基准，每10 km 一个区间将RLV 再入段划分为3 个飞行阶段，并分别进行了气动参数辨识．首先，将RLV 飞行器再入段的动力学模型转换为非线性系统的状态空间模型；其次，基于状态空间模型，将飞行器的原始状态向量进行扩维，得到由待辨识气动参数和原始状态向量组成的新扩维状态向量；然后，采用EKF算法对RLV 气动模型的扩维状态向量进行辨识，达到滤除噪声和估计未知气动参数的目的；之后，为了降低测量和过程噪声统计特性的设置对EKF 辨识结果带来的影响，在EKF 算法前向滤波和Rauch-Tung-Striebel(RTS)后向平滑过程的基础上，采用EM 算法对EKF 的测量和过程噪声的先验统计数据进行估计，基于估计所得到的精确噪声特性，能够更好地提高EKF 算法对气动参数估计的精度；最后，通过基于EM-EKF 算法与极大似然方法的气动参数辨识值对各种气动系数影响的仿真对比，验证了EM-EKF结合辨识算法的准确性．\r
\r
Abstract_English:\rAerodynamic parameters of reusable launch vehicles (RLVs) during reentry process are characterized by uncertainties and fast time-varying characteristics，resulting in strongly coupled and nonlinear aerodynamic characteristics of the RLV. Owing to these properties，the aerodynamic model is difficult to design and control，which in turn，reduces the stability of the aircraft. In the presence of system noise，the RLV reentry aerodynamic parameter identification method was proposed herein for the dynamic model of RLV using combined expectation maximization(EM) algorithm and extended Kalman filter (EKF). The reentry process was divided into three flight phases every 10 km based on the flight altitude and the angle of attack. The aerodynamic parameters were separately identified. First，the dynamic model of the reentry process of the RLV was transformed into the state space model of the nonlinear system. Second，the unknown aerodynamic parameters to be identified and original state vector were combined based on the state-space model to expand the original state vector of the aircraft to a new extended state vector. Then，the extended state vector was identified using the EKF algorithm to filter noise and estimate the unknown aerodynamic parameters. The setting of measurement and process noise statistical characteristics was reduced based on the forward filtering of the EKF algorithm and Rauch-Tung-Striebel (RTS) smoother with backward smoothing. EM algorithm was used to estimate a priori statistics that describe the measure and process noise of the EKF. This procedure can improve the accuracy of the EKF algorithm for estimating the aerodynamic parameters. Finally，the accuracy of the combined EM-EKF identification algorithm was verified via the simulation comparison of aerodynamic parameters based on the EM-EKF algorithm and maximum likelihood method.\r
\r
Keyword_Chinese:可重复使用运载器；扩展卡尔曼滤波；最大期望算法；RTS平滑器；参数估计\r

Keywords_English:reusable lauch vehicle(RLV)；extended Kalman filter(EKF)；expectation maximization(EM)；Rauch-Tung-Striebel(RTS) smoother；parameter estimation\r


PDF全文下载地址:http://xbzrb.tju.edu.cn/#/digest?ArticleID=6378