空间微重力地面模拟装置基于神经滑模的自适应补偿控制  

Adaptive Compensation Control for Ground Simulation Equipment of Space Microgravity Environment Base on Neural Sliding-model

作　　者：张文辉[1,2] 叶晓平[1] 季晓明[1] 

机构地区：[1]丽水学院工学院,丽水323000 [2]哈尔滨工业大学航天学院,哈尔滨150001

出　　处：《机械工程学报》2014年第15期68-72,共5页Journal of Mechanical Engineering

基　　金：国家科技支撑计划(2013BAC16B02);浙江省自然科学基金(LZ12F02001;Y14F030011);浙江省公益技术(2013C3110);浙江省教育厅(Y201330000)资助项目

摘　　要：针对目前做复杂运动的大中型飞行器地面微重力环境难以实现的问题,提出了新颖的模拟装置,并设计了基于神经滑模的自适应补偿控制方法。该装置采用机械传动、电动机驱动和气悬浮的组合方式来实现做复杂运动的三维模拟。采用气悬浮来被动实现平面二维随动,采用力反馈的主动控制补偿技术来实现竖直方向升降。在考虑到精确模型无法获得及存在摩擦干扰的情况下,神经滑模控制策略能够保证控制系统的鲁棒性,设计的自适应神经网络控制器能够在线学习未知不确定上界,降低滑模'抖振'。仿真表明了所设计的试验系统能够满足三维运动要求,达到了较高模拟精度,对于复杂空间运动的飞行器微重力模拟具有重要工程价值。Difficult problems of simulation the space microgravity environment for big and middle experimental objects with complex move are considered. A novel system is proposed and the adaptive compensation control method based on neural sliding mode is designed. The equipments combine machine transmission, motor drive and air-bearing to realize three-dimension spatial microgravity simulation. Force feedback controller counteracts gravity real timely as controller. The scheme dispenses with accurate dynamical model. Friction non-linearity of machine system and outside disturbance are considered. Controller is designed based on state feedback. Sliding model controller is employed to compensate uncertainty model to guarantee robust. RBF neural network is used to adaptive study uncertainty up bound to reduce “chattering”. The stability is proved based on Lyapunov theory. The simulation shows that the design of test system can achieve higher simulation accuracy, has important engineering value.

关 键 词：三维微重力 地面模拟系统 神经网络 滑模控制 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]