An Intelligent Compensation Strategy for Permanent Magnet Linear Synchronous Motors Drive

An recurrent fuzzy neural network (RFNN) compensation strategy combined with the backstepping approach for permanent magnet linear synchronous motors (PMLSMs) drive is proposed. Considering the lumped uncertainties with parameter variations and external disturbance for an actual PMLSM drives, a backstepping control law is derived by backstepping design technique. However, the upper bound of the lumped uncertainty is difficult to obtain in advance in practical applications. A RFNN observer is proposed to adapt the value of lumped uncertainty, the online parameters training of the RFNN is derived by the gradient descent method. Simulated results show that the proposed method possesses better position tracking performances and robustness to lumped uncertainty for PMLSM drive.