摘要
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。
A control strategy of the triple inverted pendulum based on a luzzy neural network ann Feedforward Compensation was constructed to solve the rule number explosion in multi-variable systems. This control strategy is applicable to the muhi-variable system that its state variables can be distinguished between the kind and importance. It has not only reduced the rule numbers of fuzzy neural network sharply and made for use of expert experience, but also has good robustness and strong nonlinear adaptive ability.
出处
《模糊系统与数学》
CSCD
北大核心
2011年第4期129-136,共8页
Fuzzy Systems and Mathematics
基金
国家自然科学基金资助项目(50675186)
