Abstract
Digital twins have emerged as powerful tools for urban planning, enabling real-time simulation and monitoring of city systems. However, most existing approaches lack meaningful human participation, limiting their effectiveness for co-design processes that require stakeholder input. This article proposes a human-in-the-loop (HITL) digital twin framework for urban co-design, integrating citizen feedback, expert judgment, and automated simulations. The framework is built on three layers: a data layer for multi-source urban data, a simulation layer for predictive modeling, and a participation layer for stakeholder interaction. We implement a prototype using a mid-sized city case study, incorporating 120 participants in iterative design workshops. Results show that HITL digital twins improve design satisfaction by 34% compared to baseline simulations, while reducing conflict resolution time by 28%. Regression analysis reveals that transparency of decision algorithms and ease of interaction are key predictors of stakeholder trust. The study demonstrates that embedding human feedback loops into digital twin architectures enhances both process legitimacy and outcome quality. We discuss implications for urban governance, ethical considerations, and scalability to larger metropolitan areas.