On designing the control vectors in simplex-type sliding mode control systems

Up till now, the selection of the magnitude and direction of the control vectors in a simplex-type sliding mode control system has been based on trial and error. Typically, the control vector directions would be arbitrarily picked first and then the control vector magnitudes would be iteratively adjusted, based on simulations, to be as large as needed to demonstrate reaching of the desired sliding surface from some set of initial states. The main contribution of this paper is to remove most of the guesswork from the control vector selection process.