A Real-Time Implementation of an MPC-Based Motion Cueing Strategy with Time-Varying Prediction

Dynamic driving simulators are seeing an increasing interest in the automotive community, both in the research and industrial fields. Different aspects are involved, from virtual prototyping to rehab: racing applications are of particular interest, where simulators are exploited to improve the driver´s capabilities and test different vehicle set-ups avoiding the costs of testing on real tracks. The ability of the platform to reproduce as faithfully as possible the driving feelings is crucial in such context: this is the task of Motion Cueing Algorithms. Recently, innovative, MPC-based approaches have been proposed which improve the performance with respect to the classical, filtering-based procedures. The criticality in such approaches is the prediction phase, both for the derivation of reliable reference signals and the computational burden associated with the solution of the optimal problem. In this paper, a first strategy is proposed to derive an affordable reference by exploiting the repetitive pattern typical of the racing context, while a move-blocking approach is integrated to assure real-time capabilities.