一种RBF神经网络的直接自适应滑模轨迹跟踪控制设计  

A Direct Adaptive Sliding Mode Trajectory Tracking Control Based on RBF Neural Networks

作　　者：王文娟 李俊 WANG Wen-juan;LI Jun(Jiangxi Institute of Applied Science and Technology,College of Artificial Intelligence,Jiangxi Nanchang330100,China)

机构地区：[1]江西应用科技学院人工智能学院,江西南昌330100

出　　处：《机械设计与制造》2020年第11期183-187,共5页Machinery Design & Manufacture

基　　金：江西省教育厅科学技术研究基金项目(GJJ151318)。

摘　　要：针对存在随机干扰和参数摄动情况下的鲁棒轨迹跟踪问题,基于径向基函数(RBF)神经网络直接自适应控制与滑模控制设计了一种高精度的控制律。以战术导弹中制导段的运动模型为研究对象,将滑模变结构的策略应用到非线性控制律的设计中,针对滑模控制不连续的开关特性造成的抖振现象,利用RBF神经网络对非线性系统的自适应逼近来降低切换增益,削弱抖振。仿真验证了所设计的控制律在随机风干扰与气动参数摄动的条件下,能够实现对轨迹的精确跟踪。Aiming at robust trajectory tracking problem with random disturbance and parameter perturbation,a high precision control law is designed based on a direct adaptive radial basis function(RBF)neural network and sliding mode control.The sliding mode structure is applied to the design of the nonlinear control law based on the motion model of the midcourse guidance phase of tactical missile.Aiming at the chattering phenomenon caused by the discontinuous switching characteristics of sliding mode control,a simple RBF neural network is used to reduce the switching gain and restrain the chattering through the adaptive approximation of the RBF network to the nonlinear system.The designed control law is verified that it can accurately track the trajectory under the condition of random wind disturbance and aerodynamic parameter perturbation.

关 键 词：RBF神经网络 滑模控制 轨迹跟踪 自适应 

分 类 号：TH16[机械工程—机械制造及自动化] U462.1[机械工程—车辆工程]