Lightweight Digital Twin Frameworks for Intelligent Edge-Based Monitoring of Smart Cities: A Comprehensive Review

Abstract

Smart city infrastructures increasingly rely on Cyber-Physical Systems (CPS) and Internet of Things (IoT) nodes to enable real-time monitoring of urban mobility, environmental health, structural safety, and energy distribution. Traditional Digital Twin (DT) architectures depend heavily on centralized cloud processing, incurring severe communication latencies, high bandwidth overheads, and potential privacy vulnerabilities. To overcome these bottlenecks, Lightweight Digital Twin (LDT) frameworks deployed at the network edge have emerged as a critical paradigm shift. This paper presents a comprehensive, state-of-the-art review on lightweight digital twin frameworks designed for intelligent edge-based monitoring in smart cities. We systematically examine lightweight synchronization protocols, resource constrained containerization (e.g., K3s, Docker Micro), edge-native AI inference engines (TinyML, ONNX Runtime), and data reduction mechanisms that enable low-latency state representation. Furthermore, we formulate a taxonomy of LDT architectures, evaluate comparative performance across real-time execution metrics, identify critical research gaps regarding synchronization consistency vs. resource overhead, and present actionable insights for future research directions in resilient urban intelligence.

KEYWORDS: Digital Twin, Lightweight Frameworks, Edge Computing, Smart Cities, IoT Monitoring, TinyML, Low-Latency Synchronization.