Neuro-fuzzy state space controller for drive with elastic joint

This paper presents adaptive control structure applied for electrical drive with elastic coupling between motor and load machine. The main speed controller is based on neuro-fuzzy model. Proposed control structure works using all state variables of the two-mass system. Feedbacks coefficients have been calculated analytically. Information about state variables is given by the Luenberger observer. For better dynamics of estimation and noises reduction, implemented algorithm is modified by introduction of neural network. Concluding, on-line adaptation is applied inside of controller and observer. Simulations are confirmed using dSPACE1103 card.