摘要
模糊神经网络控制器是一种将模糊逻辑与神经网络相结合的智能控制器,其既不依赖于被控对象精确的数学模型,又能根据被控对象参数和环境的变化自适应地调节控制规则和隶属函数参数,但是存在着收敛速度慢,较多局部极小的情况下很容易陷入局部极小值等缺点。针对存在的问题,提出一种模糊神经网络控制器的优化方法。隶属度函数的参数具有全局性,用遗传算法来优化;神经网络的权值代表模糊系统的控制规则,它用神经网络的误差反传算法(BP)来调整。将算法用于航空发动机控制,实现对低压转子转速的无静差控制,与应用BP算法的模糊神经控制相比,控制性能改善较大,结果令人满意。
Fuzzy neural controller is a kind of intelligent controller which combines fuzzy logic with neural net- work. It does not require accurate model of plant and is able to learn to control adaptively by updating the parameters of the controller. Nevertheless, it has some shortcomings, such as low convergence speed, easy to fall into the local minimum points when there are lots of local minimum points. A method for optimizing fuzzy neural controller is pro- posed. The parameters of membership functions are global, and can be optimized with GA. The linking weights of neural network stand for control rules of fuzzy system, and will be adjusted with BP. The algorithm is applied to aero - engine control. It realizes the control of the rotating speed of low - pressure rotor without static error. Compared with general fuzzy neural network control system, its performance is improved greatly, and the result is satisfactory.
出处
《计算机仿真》
CSCD
北大核心
2009年第10期65-68,共4页
Computer Simulation
关键词
遗传算法
模糊控制
神经网络
Genetic algorithm
Fuzzy control
Neural network
