Neuro-sliding mode control for magnetic levitation systems

It is well known that sliding mode control (SMC) is capable of tackling systems with uncertainties. However, the discontinuous control signal causes the significant problem of chattering. Furthermore, thorough knowledge of the plant dynamics may be unknown or difficult to obtain, which makes it difficult to calculate the control law. A synergistic combination of neural network (NN) and SMC methodology is proposed. The network weights are adjusted using a modified online error backpropagation algorithm. Moreover, a new and simple approach is utilized to construct corrective controls of SMC to overcome the chattering problem. As a result, chattering is eliminated and the error performance of SMC is also improved. Experimental studies carried out on a magnetic levitation system are presented.