Intelligent Gain-scheduling Control for Multivariable Discrete Linear Time Varying Systems

This paper proposes an Intelligent Linear Parameter Varying (ILPV) control approach for multivariable discrete Linear Time Varying (LTV) systems. An optimal linear quadratic closed loop control law is developed for each identified multivariable model so that the controlled system tracks a desired trajectory over the entire time interval. A gain scheduling adaptive control scheme based on neural networks is designed, by backpropagation algorithm, to tune on-line the optimal controllers.Simulation results are shown to demonstrate the efficiency of the proposed methodology for a multivariable plant with time varying eigenvalues, important in spacecraft and robot arms control applications, for instance.