CAVIA: Enabling the Cloud-to-Autonomous-Vehicles Continuum for Future Industrial Applications

Autonomous Vehicle (AV for short) technology is expected to play a crucial role to automate industrial processes in many contexts ranging from logistics to production. Their introduction will reduce repetitive and dangerous tasks for workers, thus improving their overall health and working conditions.

AVs require an infrastructure to support management services that control AVs decisions, monitor task execution and supervise safety conditions and immediately intervene if required, e.g., for risk prevention. When multiple AVs are employed, management services can be deployed also for coordination, e.g., to manage an AV fleet as a whole.

Some of such management services will require high efficiency and very low latency to ensure timed and reliable data collection. To this aim, the role of the computing and networking infrastructure is central in order to support such requirements on a large scale. Mass adoption of AVs, however, will represent a major challenge for the existing infrastructure: AVs include a large set of sensors from ambient to detection sensors; each sensor generates an uninterrupted stream of data that is directed to management services deployed in the cloud-to-edge continuum.

In this project, we aim at developing an ecosystem of solutions to enable the next generation of the cloud-to-autonomous-vehicle continuum.

The project will produce innovative solutions in two areas:

• network management solutions to ensure efficient data collection with mobility support for reliable and timed communication between AVs and the infrastructure;
• models and algorithms for service composition and deployment with reconfiguration of network and computing resource allocation with the aim to ensure efficient orchestration of management services in the cloud-to-vehicle continuum by considering service requirements in terms of latency and reliability.

The project overall will adopt a joint cross-layer / cross-tier approach that will be employed to design all the components of the CAVIA ecosystem, a set of tightly integrated solutions that aim at managing all the aspects of the infrastructure that support AVs, including service composition, orchestration and network management.

The holistic approach adopted in CAVIA will allow to optimize the initial configuration of the infrastructure and to support proper reconfiguration in case of context or infrastructure status changes.

With respect to existing solutions, CAVIA will be the first one to provide an integrated ecosystem where network management functionalities and service orchestration/deployment are performed jointly, with specific support for mobility by design.