一种改进的单神经元PID控制策略

An improved control strategy of single neuron PID

针对传统单神经元自适应比例-积分-微分(PID)控制中比例系数调试过程繁琐冗长的问题,提出一种改进的单神经元自适应PID控制策略.采用非线性变速控制算法对比例系数K进行在线自适应调节,使K值的变化根据输入误差的大小自动调整,且调整的速度与误差大小成正比.采用设计空间搜索规则得出该算法的最优解因子.将该方法运用于直流无刷电机的速度调节,实验仿真证明,该改进的算法相对于传统PID控制和单神经元PID控制不仅收敛速度更快,而且能有效解决电机负载扰动带来的不稳定性,尤其是在大负载扰动下体现出优异的鲁棒性,极大地提升速度调节器的品质.

Since the traditional cumbersome adjusting method that the constant can just be changed repeatedly and artificially,an improved single neuron adaptive PID control strategy,which adopted nonlinear algorithm for K,was presented.K and its changing rate were designed to adjust adaptively according to the error.Since the maximum and minimum value of K and their corresponding errors were quite important factors that influence the efficiency of this new algorithm,the design space searching rule was applied to obtain the optimal factors.By modeling the speed regulation of brushless DC motor with MATLAB/SIMULINK and using this improved strategy on the control model,the results of experiment prove that it can not only speed up the convergence rate,but also effectively settle the instability caused by the load disturbance comparing with traditional PID controller and original single neuron PID controller.Particularly,the robustness and quality of this improved controller are well verified by much smoother velocity response on condition of large load.