舰载机着舰纵向时变风险权值矩阵预测控制  

Model predictive control of the longitudinal time-varying risk weight matrix for a carrier-based aircraft landing

作　　者：朱齐丹[1] 王立鹏[1,2] 张智[1] 姜星伟 

机构地区：[1]哈尔滨工程大学自动化学院,黑龙江哈尔滨150001 [2]哈尔滨工程大学科学技术研究院,黑龙江哈尔滨150001 [3]中国航空工业无线电电子研究所,上海200241

出　　处：《哈尔滨工程大学学报》2016年第11期1532-1538,共7页Journal of Harbin Engineering University

基　　金：国家自然科学基金项目(51409053;61603110);国家国际科技合作专项项目(2013DFR10030);中央高校基本科研业务专项资金项目(HEUCF160409)

摘　　要：为实现舰载机纵向自动着舰,提出时变风险权值矩阵的预测控制方法来构建舰载机纵向自动着舰引导律。首先,建立基于偏差形式的舰载机纵向着舰非线性模型,根据航母实时运动状态动态求解飞机着舰过程平衡点,并获得线性模型的动态系统矩阵;其次,提出纵向着舰高维风险建模理论,并通过BP神经网络训练样本数据建立风险模型,根据高维风险模型构建预测控制的时变权值矩阵,给出并证明求解最优控制量的若干定理,推导出相应线性矩阵不等式,并设计状态观测器来观测当前无法直接测量的着舰状态。最后,在半物理仿真平台上验证建立的着舰引导律,通过仿真曲线证明了算法的可行性和有效性。m In this paper, we propose a model predictive control （ MPC） with a time-varying risk weight matrix to achieve an automatic longitudinal landing of a carrier-based aircraft and to establish a guidance law for its longitudi-nal automatic landing. First, a nonlinear model of the longitudinal aircraft landing is established on the basis of deviations. A balance point of the aircraft is solved in the landing process according to the real-time motion states of a carrier to obtain the dynamic system matrix of a linear model. Second, a type of longitudinal landing, high-dimensional risk modelling theory is proposed. A risk model is also built by training sample data of a back propagation neural network. The time-varying matrix of the MPC is constructed using the high-dimensional risk model. Several theorems and corresponding linear matrix inequalities are listed and proved to solve the optimal control inputs. Furthermore, a state observer is designed to predict the current landing states that are unmeasured in the actual situation. Third, the automatic landing guidance law is verified on a semiphysical simulation platform. The feasibility and effectiveness of the algorithm are demonstrated by a simulative curve.

关 键 词：自动着舰 预测控制 甲板运动补偿 动平衡点 风险建模 线性矩阵不等式 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]