From Agent-Based Simulation to an IoT-Enabled Prototype for Infrastructure-Less Cooperative Parking Guidance

Abstract

Urban parking search remains a major source of congestion, travel delay, and unnecessary emissions, especially in areas where free on-street parking is not instrumented by dedicated infrastructure. This article investigates an infrastructure-less cooperative parking-guidance approach in which participating vehicles contribute lightweight parking-related evidence that is aggregated into shared heatmaps. The proposed model formalises parking-event semantics, heatmap representation, evidence updates, and a score-based parking-selection strategy that balances parking opportunity against walking distance. A SUMO-based simulation campaign evaluates the model under different user-preference parameters, system-adoption rates, and levels of initially available information. The results show that shared parking evidence can reduce search time under the evaluated conditions, including scenarios with partial adoption, while also highlighting the influence of user preferences and cold-start information availability on system performance. To examine the implementation feasibility of the information-generation mechanism assumed by the model, the article additionally presents a compact proof-of-concept pipeline based on a Bluetooth Low Energy beacon, a smartphone application, and a lightweight back-end service. Functional testing demonstrates that parking-related state-change events can be generated, transmitted, stored, spatially aggregated, and visualised without dedicated roadside parking sensors. This prototype serves only as an implementation-feasibility demonstration and does not constitute field validation of the guidance strategy or establish real-world accuracy, latency, reliability, or deployment readiness.