A Parameterized NMPC Scheme for Embedded Control of Semi-active Suspension System

Abstract

This paper proposes a parameterized NMPC scheme for control of semi-active suspension system for a quarter car vehicle model. The proposed controller is compared against a linearization based MPC to verify its performance under real-time (RT) embedded conditions. The linearization based MPC method linearizes the non-linear dynamical system/constraints at the current operating point and a linear MPC problem is solved by means of a quadratic program (QP) solver to obtain the optimal control input. The proposed parameterized NMPC method finitely parameterizes the control input and for each parameterized input, the nonlinear dynamical system is numerically simulated and the optimal input is elicited from the simulations which minimizes the objective function and satisfies the system constraints. The methods were considered on the basis of RT implementable scenario and practical viability for the target system. The methods were successfully tested by means of hardware in the loop (HIL) simulations under RT conditions for given control objective on dSPACE MicroAutoBox II hardware. The results from HIL simulations exudes better performance of parameterized NMPC against linearization based MPC in terms of computational efficiency and RT feasibility of the controller.