Enhanced model free adaptive control by integrating with lazy learning

An enhanced model free adaptive control (EMFAC) algorithm is presented for a class of nonlinear system by integrating lazy learning technique to achieve adaptive adjustment of penalty parameter. In proposed EMFAC, penalty parameter is queried from database which consists of the historical information vector, rather than set by try and error method. The main feature of proposed method is that, controller is designed merely by the measured input-output data of controlled system without any explicit or implicit use of the physical model. Its effectiveness is further verified by simulation results.