自适应反推神经网络的转台鲁棒控制器设计

针对未知非线性、外界干扰和参数摄动等不确定因素对实际转台控制系统的影响,提出了自适应反推神经网络的转台鲁棒控制器设计.首先给出自适应Backstepping控制器的设计方法及步骤,接着采用RBF神经网络对转台对象参数的不确定因素上界值加以辨识,实现转台系统的鲁棒控制.其中Backstepping鲁棒控制作为主控制器,RBF神经网络实现了不确定上界值的在线辨识.仿真结果表明,自适应Backstepping神经网络控制很好地克服了对象的不确定性,实现控制系统的较强鲁棒性,适于高精度飞行仿真转台系统的实时鲁棒控制.

Design of Robust Controller Based on Adaptive Backstepping and Neural Network for Flight Simulator

Considering unknown nonlinearities, external disturbances and parameter perturbations in the practical flight simulator system, a robust controller based on adaptive backstepping and Neural Network method is proposed. Firstly the design and reasoning process of robust adaptive backstepping technique are proposed. Secondly RBF neural network used for acquiring the upper bound of variations of system model is presented in detail. Finally a robust controller of flight simulator is realized. In the robust controller, adaptive backstepping is the main controller and RBF network controller is used to compute the value of upper bound of control system. Simulation result shows that the method is perfect and is suitable for robust control of high - precision servo system.