An improved B-spline fuzzy-neural network controller

The paper proposes an improved B-spline networks in order to increase the ability of rejecting the disturbance when it is as a fuzzy-neural network controller in real applications. By pointing out the shortage existed in the conventional B-spline function as the fuzzy membership function, the paper proposes the relevant improved method in such way as both to satisfy B-spline features, and also avoid system unstable factors. The actual experiments for a DC motor speed control system verify the effectiveness and practicability. The paper gives out the design procedure and the experiment results of the improved B-spline networks