Enhancing Intelligent Transportation Systems with Optimal Cloud-to-Things Resource Allocation

Abstract

Urban sensing for Intelligent Transportation Systems (ITS) increasingly relies on the Cloud-to-Things Continuum (C2TC), to balance latency, reliability, and computational demand. Systems based solely on cloud platforms can no longer guarantee these requirements, especially as the number of sensors and monitoring devices continues to grow, increasing the overall data exchange. By deploying edge Artificial Intelligence (edge-AI) models for vehicle detection and counting at the network edge, intersections can adapt signal plans locally, reducing reaction times and data traffic. In this paper, we tackle the problem of optimal task placement in the C2TC to meet the temporal requirements of edge-AI–based ITS. We propose a workflow that integrates J-NECORA, an analytical framework that computes optimal resource allocation under Quality of Service (QoS) constraints, with the EdgeCloudSim simulator, to assess additional Key performance indicators (KPIs) such as network time, node utilization and missed deadlines. The workflow is evaluated in an urban traffic scenario, and results show that it provides deeper and broader insights compared to a purely analytical approach, confirming the effectiveness of the proposed approach for designing time-critical, edge-enabled ITS.