基于模糊CMAC网络的非线性自适应逆控制

针对非线性自适应逆控制中非线性对象的建模和逆建模的精确性这一问题,提出一种基于模糊小脑模型关节控制器(Fuzzy Cerebellar Model Articulation Controller, FCMAC)网络的非线性自适应逆控制方案.将模糊逻辑思想嵌入到CMAC中构成FCMAC来对非线性对象进行较精确的逆建模,从而构建逆控制系统.在对象特性未知的情况下,选用BP网络来对象进行正建模,并由BP网络的辩识结果来对FCMAC的参数进行调整.仿真实验表明了该方案的有效性,且验证了其控制效果较单纯的CMAC网络逆控制更理想.

Non-linear Adaptive Inverse Control Based on Fuzzy CMAC Network

Adaptive inverse control is a new way to design control system and tuner. Its successful application is determined by the accuracy of non-linear modeling and inverse modeling. A method of non-linear adaptive inverse control based on Fuzzy Cerebellar Model Articulation Controller (FCMAC) is proposed to solve this problem. Theoretically,FCMAC which combines the advantages of fuzzy logic and CMAC can approach any complex non-linear function in any precision. According this characteristic,an exact inverse model can be obtained,and then construct the inverse control system. Unknown the characteristic of the plant,choosing BP Network can make a modeling for the plant. Also using the result of identification of BP Network the Weight of FCMAC can be tuned. Simulating experiment proves the feasibility of the control system and that the effect of control system is better than CMAC inverse control.