Cloud-native roadside infrastructure saves energy with demand-driven V2X

Intelligent roadside infrastructure is critical for cooperative intelligent transport systems (C-ITS) that support vehicles with automated driving systems. However, as the number of roadside units grows, efficient operation becomes challenging. In a new paper accepted for IEEE ITSC 2026, Lukas Zanger and colleagues from RWTH Aachen University present a cloud-native architecture that uses a Kubernetes cluster spanning both roadside units and a cloud server. This allows demand-driven orchestration: resource-intensive services like V2X-based collective perception are dynamically deployed only when a connected vehicle is nearby. Without demand, the application remains inactive, cutting energy consumption, channel congestion, and hardware wear. The approach was validated in a real-world experiment in Aachen, demonstrating that collective perception starts in time for the vehicle to benefit.

The researchers also analyzed V2X recordings from the test field to estimate the energy-saving potential of demand-driven operation. Their results confirm the practical feasibility of this cloud-native, demand-driven approach and indicate its potential to improve scalability and efficiency in future C-ITS deployments. The paper highlights that such an architecture can reduce the operational cost of roadside infrastructure while maintaining performance. As autonomous and connected vehicles proliferate, this work provides a blueprint for building energy-efficient, scalable roadside systems that only consume resources when they are needed.