Dual-Mode Infinite Horizon Constrained Model Predictive Control Parameterized in Terms of Laguerre Polynomials

The computational cost of model predictive control is a major concern in practice. This paper aims to accelerate dual-mode model predictive control by reparameterization of the control signal. In the dual-mode control, the range of model predictive control is divided into two disjoint set; in the first set all constraints are active and a nonlinear constrained optimization problem is solved to compute the control signal. In the second set, all constraints are satisfied inherently so an unconstrained optimization problem is solved to compute the model predictive control signal. In this study, when the dynamical system is in the first set i.e. set of constraints are active, the control signal is represented by a minimal set of Laguerre polynomials. Since this parameter set is minimal, the size of the nonlinear optimization is reduced and model predictive control signal can be calculated faster. As an exemplary problem, a quarter-car active suspension control problem was studied. With numerical simulations, it has been shown that dual-mode infinite horizon constrained model predictive control employing Laguerre polynomials can respond faster than a regular model predictive control.