2021/06/15 by Tim Brüdigam, Jie Zhan, Brüdigam, Tim +5
Chemical Engineering · Engineering · #Advanced Combustion Engine Technologies #Advanced Control Systems Optimization #FOS: Electrical engineering #Real-time simulation and control systems #Systems and Control (eess.SY) #Vehicle Dynamics and Control Systems #Vehicle emissions and performance #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.2106.08463
openalex publication_date 2021/06/15 · openalex created_date 2025/10/24 · openalex updated_date 2026/07/28
Model Predictive Control (MPC) has shown to be a successful method for many\napplications that require control. Especially in the presence of prediction\nuncertainty, various types of MPC offer robust or efficient control system\nbehavior. For modeling, uncertainty is most often approximated in such a way\nthat established MPC approaches are applicable for specific uncertainty types.\nHowever, for a number of applications, especially automated vehicles,\nuncertainty in predicting the future behavior of other agents is more suitably\nmodeled by a twofold description: a high-level task uncertainty and a low-level\nexecution uncertainty of individual tasks. In this work, we present an MPC\nframework that is capable of dealing with this twofold uncertainty. A scenario\nMPC approach considers the possibility of other agents performing one of\nmultiple tasks, with an arbitrary probability distribution, while an analytic\nstochastic MPC method handles execution uncertainty within a specific task,\nbased on a Gaussian distribution. Combining both approaches allows to\nefficiently handle the twofold uncertainty structure of many applications.\nApplication of the proposed MPC method is demonstrated in an automated vehicle\nsimulation study.\n