Adaptive control of nonlinear black-box systems based on universal learning networks

This paper presents an adaptive control scheme for nonlinear black-box systems based on the use of universal learning networks (ULN). A ULN nonlinear controller is constructed in a similar way to linear stochastic control theory. In the obtained ULN controller, some node functions are known, while others are unknown. Each unknown node function is reparameterized using an adaptive fuzzy model. A robust adaptive algorithm is developed to adjust the unknown parameters in the controller. The effectiveness of the proposed control scheme is examined via numerical simulations