Digital Twins as Decision Support Systems for Sustainable Smart Cities: a Traffic Analysis Perspective

Digital Twin (DT) technology has revived the interest in the Smart City concept, allowing to build accurate digital models of complex phenomena, serving as decision-making support systems for different urban stakeholders. In this thesis, we focus on main DT building blocks allowing for efficient analysis of the urban environment. The technology could be employed as a simulation tool to analyze environmental impacts, exploring alternate scenarios. Grounding the study, we explore the use of the technology via two use-cases: Acc2Twin for transport packaging and ParallelTwin for traffic analysis — focusing on road network partitioning. The Acc2Twin cases showcases a data processing pipeline aimed at building a continuous road network (path) topology, given in input to a first-principles DT, tasked with simulating stretch forces exerted on palletized products during transportation. Acc2Twin interpolates route data to an arbitrary precision to allow for the replication of real transport conditions. The second use-case, ParallelTwin, aims at optimizing parallel simulations of vehicular traffic with the SUMO software package, reducing simulation time, and enhancing system performance without altering the core pre-existing simulation program. We report on PartitionTwin’s framework, operational mechanisms, communication protocols, and the METIS algorithm for efficient graph partitioning, researching efficient partitioning weights to obtain a balanced simulation workload.