Model-Predictive Cruise Control for Direct Teleoperated Driving Tasks

• Published in: 2021 European Control Conference (ECC) pp. 1808 - 1813
• Main Authors: Schitz, Dmitrij, Graf, Gaetano, Rieth, Dominik, Aschemann, Harald
• Format: Conference Proceeding
• Language: English
• Published: EUCA 29-06-2021
• Subjects: Europe; Hardware; Predictive models; Process control; Real-time systems; Safety; Stability analysis
• DOI: 10.23919/ECC54610.2021.9655187

Teleoperated driving, where an operator remotely takes over control for some time-span, represents an important fallback mode for autonomous vehicles. However, the given communication delay is one of the major challenges that may affect system stability. In the area of direct teleoperated driving, the vehicle trajectory prediction as the Predictive Display is the state-of-the-art technique to deal with this issue. Its effectiveness is highly dependent on the actual communication latency as well as the operator's control commands. This contribution presents, hence, an assistance system for direct teleoperated driving with extended automation functions to modify the operator's control commands in order to increase safety and comfort. Based on predicted object data, a model predictive trajectory optimization ensures collision-free longitudinal dynamics. Real driving experiments confirm the effectiveness of the proposed approach and highlight both the achieved driving safety and real-time capability.