Digital Twin and Cloud Based Remote Control of Vehicles

Cloud-based, centralized control system architectures have been published in recent years, examining the applications, structure and validity of the idea of controlling a vehicle, or multiple vehicles directly from the cloud. The literature presents possible advantages and disadvantages of Control as a Service (CaaS), as well as multiple viable applications, however no real-car real-world implementation exists to the best of our knowledge. In our experiments, we demonstrate a real-life and real-time vehicle control method that runs in the cloud and reacts adaptively to the vehicle’s surroundings. In the presented tests, acquisition of the vehicle and environment state is performed from a collective perception-based digital twin that is also produced in our cloud-based Central System using data streams from our distributed sensor network. We conduct real-world physical vehicle tests, and establish that cloud-based centralized control is a viable solution, and an interesting research direction. We test our vehicle control in one mixed reality, and two real-world scenarios on the ZalaZONE proving ground in Hungary, Zalaegerszeg. Our experiments prove that cloud-based low-level vehicle control is feasible.