A Self-Organizing Global Sliding Mode Control and Its Application to Active Power Filter

In this article, a self-organizing global sliding mode control (GSMC) is developed for a class of dynamic systems, whereby modeling uncertainties are estimated by metacognitive fuzzy-neural-network (MCFNN) framework. First, a GSMC is designed for the tracking of reference signals to eliminate the reaching mode and chattering phenomenon. To overcome the drawbacks of GSMC, the control law is designed based on MCFNN instead of the uncertain information. Distinguished from the fixed structure schemes, MCFNN can restructure the network structure and parameters by extracting useful input data not all data. Moreover, in order to alleviate redundant or inefficient computation, only the parameters of the rule nearest to the current data instead of all rules are updated online based on Lyapunov stability analysis. Finally, simulation and experimental investigations on active power filter are employed to verify the control performance of proposed controller.