论文部分内容阅读
为解决T akag i-Sugeno型模糊神经网络在控制多变量系统时的规则组合爆炸问题,提出一种误差前馈补偿的模糊神经网络控制方案,有效实现了三级倒立摆的稳定控制。该控制方案适用对状态变量可按性质和重要程度划分的多变量系统的控制,大大减少了模糊神经网络控制器的规则数,有利于利用专家的控制经验,具有良好的鲁棒性和非线性适应能力。
In order to solve the problem of regular combinatorial explosion when T Akk i-Sugeno fuzzy neural network is used to control multivariable systems, a fuzzy neural network control scheme with error feed-forward compensation is proposed to effectively control the three-stage inverted pendulum. The control scheme is applicable to the control of multivariable systems whose state variables can be divided according to their nature and importance, which greatly reduces the number of rules of the fuzzy neural network controller, which is good for utilizing expert control experience and has good robustness and nonlinearity adaptability.