基于H∞环路成形和自适应神经模糊推理系统的模糊控制器设计

H_∞环路成形方法设计的控制器阶次较高,不便于工程实现和参数调整;用传统方法确定模糊控制器隶属度函数的参数和模糊规则比较费时且难以保证鲁棒性能和时频域性能指标.针对上述情况,提出了一种综合运用H_∞环路成形和自适应神经模糊推理系统来设计模糊控制器的方法.首先采用H_∞环路成形设计方法,得到鲁棒裕量、动态和稳态性能都符合要求的控制器,然后用自适应神经模糊推理系统来逼近此控制器,最后根据自适应神经模糊推理系统参数确定相应的模糊控制器规则和参数.该方法确定模糊控制器隶属度函数的参数精确而省时,且能保证控制器具有较强的鲁棒性和良好的控制效果.通过对小车倒立摆系统进行的仿真,验证了该控制器设计方法的有效性.

Design of fuzzy controller based on H-infinity loop-shaping procedure and adaptive-nerural-network-based fuzzy interference system

Controllers designed by H-infinity loop-shaping procedure are too complex to be applied to engineering,and it is time-consuming to decide fuzzy rules and membership function's parameters of fuzzy controllers using traditional method.To overcome the shortcoming,a design of fuzzy controller is proposed by using both H-infinity loop-shaping and adaptive-nerural-network-based fuzzy interference system (ANFIS) synthetically.First,a controller was obtained which had enough stability (margin),good dynamic properties and steady-state behavior using H-infinity loop-shaping design procedure.Then,adaptive nerural-network-based fuzzy interference system was used to approximate the controller.Finally,corresponding parameters of fuzzy controller were set according to adaptive nerural-network-based fuzzy interference system parameters.It was easy to confirm the rule and accurate parametes of fuzzy controller using this method,and the controller had enough stability margin and good control quality.Simulation results were given to show the validity of the scheme.