摘要
提出了一种基于经验数据而非语言的T-S模糊控制器设计和优化方法.此方法分为三个阶段,第一阶段依据输入变量的范围来确定输入变量的高斯型隶属度函数;第二阶段在不改变输入变量隶属度函数的前提下,对经验数据施加递推最小二乘法以确定T-S模糊控制器的后件系数;第三阶段,使用梯度下降方法同时优化控制规则的前件参数和后件参数.倒立摆的仿真试验结果验证了该方法的有效性.
An approach to design and optimize T-S fuzzy logic controller was proposed based on empirical data rather than linguistic. There were three parts in it. Firstly, the Gaussian membership function of input variable was to be determined according to the range of input variable. Then, the coefficients of consequent components in T-S fuzzy logic controller were to be determined with recursive least square algorithm for the empirical data while the membership function parameters of input variable were unaltered. Finally, all the parameters of T-S fuzzy logic controller made up of antecedent parameters and consequent parameters were to be simultaneously optimized with gradient descent algorithm. The effectiveness of this approach was verified by means of simulation of inverted pendulum.
出处
《兰州理工大学学报》
CAS
北大核心
2007年第3期75-78,共4页
Journal of Lanzhou University of Technology
基金
甘肃省科技攻关项目(RG5035-A52-007-02)
关键词
T-S模糊控制器
经验数据
递推最小二乘法
梯度下降法
倒立摆
T-S fuzzy logic controller
empirical data
recursive least square algorithm
gradient descent algorithm
inverted pendulum
