机构地区: 北京科技大学计算机与通信工程学院
出 处: 《控制与决策》 2010年第6期939-942,957,共5页
摘 要: 针对一类不确定非线性系统的跟踪问题,利用神经网络和动态面技术设计控制器,提出一种控制器参数自寻优策略.在每个子系统中,应用径向基函数(RBF)神经网络逼近该子系统中的不确定项;在每一步递推中,引入一个滤波器以克服反推技术中控制项爆炸的缺点.通过定义一个优化目标函数,应用梯度法在控制器参数可行解中寻找一组最优的控制器参数.数值仿真表明该方案是可行的. For a class of uncertain nonlinear systems tracking problem,the controller is designed by using dynamic surface control method and radial base function(RBF) neural network,and the controller parameters self-optimize strategy is proposed.The dynamic surface control can overcome the control of explosion in backstepping technique by introducing a filter at each step of the recursive procedure,and the uncertainties are approximated by using RBF neural network in each subsystem.Then an optimal objective function is defined by using gradient optimization method to search a group of optimal controller in the feasible solution of the controller parameters.Finally,numerical simulation shows the feasibility of the scheme.
领 域: [自动化与计算机技术] [自动化与计算机技术]