基于线性参数神经网络的非线性系统稳定自适应控制

Stable Adaptive Control for Nonlinear Systems Based on Linear parameter neural networks

提出适用于多种网络类型的神经网络稳定自适应控制设计思想 ,在神经网络逼近误差界未知的条件下 ,对该误差界进行在线自适应估计 ,研究基于线性参数神经网络的仿射非线性系统稳定自适应控制。采用 L yapunov函数方法证明系统状态变量、网络权值矢量、网络逼近误差界的在线估计及输出跟踪误差的收敛性。仿真结果表明 ,该方案跟踪性能良好 ,稳态误差较小 ,系统输出能快速跟踪目标信号。

A novel idea to design SNNBAC (stable neural network based adaptive control) suitable for several types of neural networks is presented. Under the condition that the network approximation error is unknown,stable adaptive control for uncertain affine nonlinear systems based on linear parameter neural networks is studied.The states of the system, network weights, network approximation errors′estimations and tracking error are proved to be convergence using Lyapunov methods. The simulation results show that the static tracking errors are small and the output can rapidly track the reference signal.