基于神经网络的非线性大系统分散自适应控制

Decentralized adaptive control of large-scale nonlinear systems based on neural network

针对一类非线性大系统,提出了一种利用神经网络的分散输出反馈自适应控制器的设计方法.对于每一个非线性的有限维子系统,采用输出反馈线性化系统;然后利用基于神经网络的自适应补偿器逼近和补偿系统建模偏差、干扰及系统关联项.从而简化控制器设计,提高跟踪精度,并使系统具有一定的鲁棒性.仿真结果表明,所提出的算法简单,易于实现,能够使闭环系统具有良好的跟踪性能.

A decentralized adaptive control design is proposed based on neural network for a class of large-scale nonlinear systems. For each subsystem, employing an output feedback linearization system and adaptive compensation system to set up an error,disturbance and the interconnections with a neural network. It simplifys the design of the controller, improves the dynamic performance and makes the system robust. Simulation results shows that the algorithm is simple, efficient, and makes the system a good tracking performance.