A New Approach of Neurofuzzy-Based Control Design and Analysis Applied on an EV Electromechanical System

摘要

Electric vehicles (EVs) constitute an attractive solution for transportation caused by both environmental and economic reasons. In view of their complexity as electromechanical structures, EVs are multi-input nonlinear systems which cannot be easily controlled and analysed. Hence, in order to operate an EV in a reliable and desired manner, neurofuzzy controllers (NFCs) are proposed, implemented and examined. The latter can deal with uncertain environments, due to their capabilities in learning, adaptation and fault tolerance by incorporating a properly adapted expert knowledge. However, in the paper, the cumbersome task of proving stability of the NFC driven electromechanical system is provided by slightly modifying the adaptation dynamics needed for the learning procedure. This is the main theoretical contribution presented by this work, since it makes possible to ensure input-to-state stability for the entire system and convergence of the system states to the desired equilibrium. The conducted simulations verify the satisfactory system performance and confirm the attained theoretical results.