改进型B样条模糊神经网络

摘要

提出了一种改进型B样条模糊神经网络，用以增加B样条函数在作为模糊神经网络控制器实际应用中的抗干扰能力.通过指出常规B样条模糊隶属函数构造中存在的不足，提出了针对性的改进措施，从而既满足了B样条本身的性质，又从理论设计上避免了使系统不稳定的因素.在实际电机速度控制系统的对比实验中证明了所提改进方法的有效性和实用性.文中给出了改进型B样条网络的设计过程以及实验结果.%The paper proposes an improved B-spline fuzzy neural network, which may be used to increase the ability of resisting disturbances when it is as a fuzzy neural network control in the actual applications. The paper proposes the effective ways in order to overcome the shortages in the construction of B-spline fuzzy membership functions, which both satisfactory the properties of B-spline itself, and also avoid factors of system instability from the theory design point. The actual motor velocity control system experiment has verified the effectiveness and the practicality. The paper gives the design procedure and the experiment results of the improved B-spline networks.