A simulation-based methodology for evaluating IoT architectures in smart street lighting systems

Abstract

Urban Internet of Things (IoT) infrastructures support smart-city services by connecting assets with sensing, communication, and management systems. Their modernization requires municipalities to compare device organizations, communication architectures, and investment strategies before large-scale deployment. Such comparisons are hindered by incomplete deployment information, limited operational measurements, and fragmented technical, operational, and financial assessments. This thesis addresses this gap by developing and evaluating a simulation-based methodology for integrated assessment of urban IoT modernization under limited-data conditions. The methodology comprises three layers. Layer 1 represents assets and generates the states and data needed to simulate operational behavior. Layer 2 defines device organization, data aggregation, and transmission through centralized or decentralized architectures. Layer 3 evaluates communication feasibility and relates architectures and operations to infrastructure requirements, Capex, Opex, and life-cycle costs. Together, the layers connect system behavior, observability, technical performance, and financial implications within one assessment. To demonstrate its application in smart street lighting modernization, the proposed approach was implemented in SELIS (Smart Energy Lighting Infrastructure Simulator). SELIS simulates streetlight-level electrical and environmental telemetry, operating states, and failures; organizes assets into Panel or Edge/Fog architectures; represents telemetry flows in Real-Time and Simulation Modes; analyzes Edge/Fog collision losses; and estimates infrastructure, communication, storage, energy, and life-cycle costs. The evaluation was conducted through SELIS using two Brazilian street lighting datasets. The first was a large-scale georeferenced inventory from Brasilia, with 267,886 streetlights, used to assess clustering, architecture organization, communication behavior, infrastructure costs, data transmission, storage, and consumption scenarios. The second was a telemetry dataset from Belo Horizonte, used to validate simulated electrical behavior against measured voltage, current, power factor, and energy data. The results show that SELIS can evaluate IoT deployment architectures and quantify how architecture choice, reporting configuration, and consumption assumptions affect scalability, cost, and operational behavior. The thesis contributes a generalizable methodology for evaluating IoT architectures before large-scale modernization. By adapting Layer 1 to domain-specific operational requirements while retaining Layers 2 and 3 as common components, the methodology supports integrated communication, observability, and technical-economic assessment across IoT domains.