6DoF Stewart motion platform control using switchable model predictive control

Abstract

Due to its high rigidity, manoeuvrability, and strength-to-weight ratio, the 6-DoF Stewart platform is widely used in flight simulators for replicating pilot motion cues. However, upset prevention and recovery training (UPRT) involves rapid angular changes that exceed motor tolerance, and classical washout filter (CWF)-based motion cueing algorithms (MCAs) struggle to meet high-accuracy and fast-response requirements. This study develops a model predictive control (MPC)-based MCA to manage nonlinearities and workspace limitations in hexapod simulators. To address control uncertainties from constraint extraction (COTC), a switchable MPC (S-MPC) architecture is proposed for adaptive response. Simulations show that within the operating envelope, MPC-MCA achieves high tracking accuracy, while outside it, the S-MPC mechanism provides optimal switching control. Under horizontal stall UPRT conditions, the proposed S-MPC-MCA improves motion tracking performance by 42.34% and 65.30% over MPC-MCA and CWF-MCA, respectively, based on the average absolute scale (AAS) criterion.