Research on Dual Kernel Fuzzy Controller Algorithm Based on Multilateral Learning

ABSTRACT For affine nonlinear systems with system uncertainties, a dual kernel fuzzy control scheme based on a multilateral learning mechanism is proposed in this paper. Firstly, based on the designed dual kernel fuzzy control structure, the nonlinear approximation error is effectively reduced, and the dynamic response time is shortened. Secondly, under the effect of a multilateral learning strategy, the multilateral parallel dual kernel fuzzy controller further approximates the nonlinear part of the system, so that the approximation error can converge to a small neighborhood of zero in finite time. In addition, the reference value of model uncertainty is constructed based on the relationship between the error differential and a second‐order filter, and the parameter update law of the multilateral fuzzy controller is designed in combination with the estimated value, so as to improve the response capability and approximation accuracy of the control system to uncertainty. The saturation transfer function is introduced to solve the problem of parameter oscillation in the control process. The stability of the control system is proved by constructing a Lyapunov function. The effectiveness and advancement of the multilateral learning dual kernel fuzzy controller are verified in the inverted pendulum system and the two‐degree‐of‐freedom manipulator by MATLAB simulation platform.