Robust emerged artificial intelligence speed controller for PMSM drive

摘要

Artificial intelligence based fusion (AIF) is a new soft optimization method that is based on the emerged science of soft computing (expert system, fuzzy logic, neural Network and genetic algorithm...) with optimal mathematical state equation (Extended Kalman filter...). In this paper we propose two optimized soft to constraint the new controller for PMSM. First, we propose a recurrent neural network controller trained with extended kalman filter results show that is better than backpropagation. Second, we employ GA to solve EKF covariance optimization problems. The approach that we use does not require any additional mathematical model of the dynamical system beyond those that are required for classical automation problems. The constrained hybrid artificial controller algorithm is compared with solutions based on a conventional controller, classical recurrent neural network controller (RNNC) and genetic algorithm (GAC) the simulated results demonstrate that constrained HAIC is more suitable for modern automation.