A Novel Adaptive Learning Systems for Type-2 Fuzzy Controller Based on Elman Neural Network

Although its advantages of handling input uncertainties and noises, Type-2 Fuzzy controller faces the challenge of the trial-and-error process for evaluating so many control parameters. To overcome this issue, an Elman neural network with the adaptation laws was proposed to make a hybrid model that leverages the expressive power of Type-2 Fuzzy Logic and the predictive capability of Elman Neural Networks to capture temporal dependencies. The adaptation laws employ the gradient descent method to dynamically update network parameters in realtime. To demonstrate the effectiveness and practical utility of the proposed control system, numerical simulation results are presented, specifically on its performance in trajectory tracking control for cart inverted pendulum systems.