自适应神经元控制系统的理论分析

A Theoretical Analysis of an Adaptive Neuron Control System

对自适应神经元控制系统进行了研究,得出了关于系统的控制作用、稳定性定理的理论结果,为这种新颖的控制系统的工程应用以及神经元控制器件的研制提供了理论依据.

The control function and stability of an adaptive neuron control system are studied. The results show that a system structure with neuron control consists in that the control signal produced by the neuron is composed of the feedforward proportional control, feedback proportional control and feedback differential control. The neuron control is a multilayered and multimodal control structure with the feedback control and feedforward control, open loop control and closed loop control integrated into a single whole. They are mutually associative and compensative. The theorem and corollary of stability of neuron control systems are given. By the very simple algorithm proposed for the neuron controller, quick adjustment and tracking can be realized without any overshooting for a complicated and changeable but stable controlled object. No integral function is needed to eliminate the static error. With very good adaptivity and robustness, the neuron control system may also be useful for certain unstable objects.