Federated digital twins platform for smart city logistics: A knowledge-driven approach

Abstract

Urban logistics faces increasing pressure from rising population densities, escalating delivery demands, and constrained urban resources. Traditional logistics systems struggle to adapt to real-time urban dynamics, leading to inefficiencies, congestion, and environmental concerns. A key challenge lies in mobilizing underutilized assets, such as off-hour freight parking, and adopting multimodal solutions to navigate diverse and increasingly strict regulations, thereby enhancing both sustainability and operational efficiency. However, effective management and utilization of these assets require real-time visibility, cross-stakeholder collaboration, and intelligent decision-making. This study proposes a federated digital twin platform to enhance logistics operations efficiency by integrating asset management and knowledge-driven operations management, relying on real-time asset visibility and delivery knowledge, such as destination characteristics and preferred logistics modalities. Unlike traditional logistics planning, which relies on static assumptions, our approach adapts to urban constraints by continuously querying real-time asset information and integrating logistics-related knowledge into operations management. To assess the effectiveness of this approach, an optimization-based simulation framework with decision-making tools is developed. The study evaluates multi-echelon logistics networks, incorporating micro-hubs, dynamic transshipment points, and multimodal logistics options, including on-foot porters, E-cargo bikes, and Road Autonomous Delivery Robots (RADRs). Findings demonstrate that integrating federated digital twins with knowledge-driven approaches, such as destination-based clustering and modality selection, reduces costs by over 50% and emissions by more than 30%. This study underscores the transformative potential of digital twins in enabling real-time, knowledge-driven operations management, and fostering more sustainable and efficient urban logistics systems. • Federated Digital Twins enhance resource visibility and knowledge-driven logistics. • Dynamic asset management minimizes waste, emissions, and inefficiencies. • Constrained K-Means optimizes modality selection using destination-based clustering. • Simulation evaluates rule-based decision-making for adaptive urban logistics. • Findings guide LSPs and policymakers in optimizing logistics operations and policies.