Parallel Model Predictive Control with feedback compensation

In this paper, we investigate utility of parallel processing in control system design. Control engineering has been developed according to development of computer systems. In recent years, computational speed has been increased by using parallel processing, and we need control method that takes advantage of the parallel processing. For parallel processing, we propose a feedback compensation method for Model Predictive Control (MPC), and a control method using unscented transformation for stochastic systems. To confirm the effectiveness of these proposed methods, numerical simulations are demonstrated. Simulation results show that the proposed method is faster than the conventional method. It is also shown that mean square error is smaller than the conventional method using the stochastic control when the system is disturbed by measurement noises.