Neuroadaptive variable structure control of mass transit trains

With the rapid development of mass transit system, how to improve the control accuracy of the train become increasingly important. During the train's running process, the resistance force is changing with some factors, such as, speed and track geometry, and these factors play an important role in tracking accuracy. Most existing methods assume the resistance force is available for feedback control or consider constant resistance coefficients. Contrasted to these methods, a neuroadaptive variable structure controller for automatic train speed and position tracking under varying operation conditions is proposed in this paper. We consider the case that the basic resistance forces and additional resistance forces are both time-varying and unknown. This method is proposed to achieve high precision position and speed tracking. The fundamental principle of this method is to design the control using combination of neural network and adaptive variable structure technique. © 2011 AACC American Automatic Control Council.