一类大规模系统神经网络自适应控制方案设计

针对一类大规模非线性机械系统，如自动化高速公路汽车空位调节和连续搅拌反应釜等系统，设计了一套基于前向神经网络的分散自适应控制方案以实现对其有效控制．作为直接自适应控制器，神经网络被用于逼近未知函数，所设计的两个鲁棒控制项分别用于消除互连效应和干扰项．系统稳定性得到严格证明，通过仿真进一步验证了方案的有效性．

Adaptive control of a class of large-scale systems using neural networks

For a class of large-scale mechanical nonlinear systems, such as inter-vehicle spac- ing regulation in a platoon of automated highway systems and continuous stirred tank reac- tion （CSTR）, this paper designs a neural network （NN-based） decentralized adaptive control scheme to control the systems effectively. As a direct adaptive controller, neural network is used to the approximately unknown function of subsystem. Two designed robust control terms are respectively applied to dispel interconnection effect and disturbances. The stability is proved rigidly. The effectiveness is validated through simulation.