Nonlinear Predictive Controller Design Based on Subspace Identification And Bilinear Modeling

In this paper, Model Predictive Control (MPC) extended to a subclass of non-linear systems based on the subspace approach using the bilinear model. The merit of this approach is that MPC controller design is obtained directly from subspace matrices, which are identified based on system input/output data. This method is useful for unknown systems. Linear controllers cannot describe the practical systems because the practical systems are non-linear, so we proposed the MPC controller design based on the bilinear subspace model.