Extreme learning machine-based stable adaptive control for a class of nonlinear system

Extreme Learning Machine (ELM), recently developed by Huang et al., has been demonstrating an exciting learning algorithm for Single hidden Layer Feedback Neural Networks (SLFN). In this paper, the ELM has been introduced to approximate the unknown functions, which may not be parameterized and so make it impossible to develop an adaptive controller. Besides, the Nussbaum-type gain method is also incorporated into the controller design to counteract the unknown coefficient of the control section. It is proved that the proposed approach is able to ensure boundedness of all the signals in the closed-loop system, and the state variables converge to zero asymptotically.