不确定性的非线性系统神经网络L_2增益控制

Neural network L_2 gain controller for a nonlinear system with uncertainty

针对具有不确定性的仿射非线性系统,设计了神经网络L2增益控制器,使得控制系统为有限增益L2稳定的.利用Fourier神经网络的函数拟合能力,给出了满足HJI不等式的存储函数的一般结构,并利用遗传算法对神经网络权系数进行优化,设计相应的神经网络L2增益抗干扰控制器,使得闭环系统满足相应的L2性能准则.对于L2输入信号,要求控制系统设计为使得输入-输出映射为有限增益L2稳定的并有尽量小的L2增益参数.针对搅拌式化学反应器控制实例,通过数字仿真,证明此方法能够达到预期的L2性能准则.

A neural network controller with L2-gain was developed for an affine nonlinear system with parameter uncertainty. The controller stabilizes the closed-loop control system with a finite L2-gain. The general structure of the storage function was formulated based on a Fourier neural network system＇s fitting capacity, which was satisfied by a Hamihonian-Jacobi inequality （HJI）. Moreover, by employing the optimization of a genetic algorithm to the weighting of the neural network system, the neural network system with its anti-disturbance system was able to meet the criteria of L2-gain performance. For an L2-gain input signal, the closed-loop control system needs to stabilize finite L2 gain to input-output mapping and have the parameters of L2 gain as small as possible. In a stirred-tank chemical reactor control example, simulation results demonstrated that the proposed method is feasible and can meet the criteria of L2-gain performance.