【摘要】 针对传统比例-积分-微分(PID)控制器参数整定困难的问题,提出一种新型比例-积分(PI)参数自动寻优控制器(SPO).先确定PI参数的取值范围,按一定取值间距建立PI参数的查找表;针对反馈跟踪控制的3个性能指标,包括上升时间、超调量和稳态误差,分别根据查找表的参数进行轮询式测试,建立3个独立的性能评价数据库;根据满意度函数映射和自动规则产生算法,将3个评价数据库分别用函数表示;为了综合考虑3个性能指标,使系统控制的整体性能达到最佳,引入了价格函数;最后通过基于层截面颗粒度分析的图形算法,寻找出最优的PI参数.在原子力显微镜(AFM)上的实验证明,本控制器能自动寻找出适合当前系统状态的最优PI参数,可避免由于人为经验因素对扫描成像造成的影响,从而取得更好的扫描图像.更多还原

【Abstract】 To deal with the problem of tuning difficulty of proportion-integration-derivation(PID) controller parameters,scanning parameters optimization(SPO),a new controller for parameters tuning of proportion-integration(PI) controller was proposed.SPO determined the parameters value range first,and made a lookup table for PI parameter groups which contained both proportional action coefficient and integral action coefficient.It then scanned the lookup table by testing the parameters group about their performance indices,such as rise time,overshot and steady state error to make three independent databases for the performance evaluation.Satisfactory degree function(SDF) and automatic rules generating algorithm(ARGA) were developed within this controller to represent the three evaluation databases with functions.Cost function was introduced to comprehensive consideration of the three performance indices to optimize the performance of the whole system.Based on cross section granularity analysis method,the optimal parameter group was obtained.Experiment with the control system on atomic force microscopy(AFM) shows that the controller can avoid influences caused by anthropogenic factors,and identify the optimal PI parameters for the current system state so that better scanning results can be achieved.更多还原