Can Digital Twins Keep Cities Moving When Transport Systems Fail?

Cities are increasingly turning to digital twins to keep transport systems moving through congestion, infrastructure failures, extreme weather, cyber incidents and service disruptions. However, a major new review published in the MDPI journal Smart Cities warns that the technology's promise is running ahead of the evidence.

The study, "Digital Twins in Intelligent Transport Systems: A Systematic Review of Resilience Mechanisms and Smart City Implications," was authored by Badr Machkour, Naoufal Rouky, Ahmed Abriane, Mouhsene Fri and Othmane Benmoussa. Reviewing 61 studies published between 2020 and 2025, the authors find that digital twins are evolving from virtual replicas into decision-support systems capable of monitoring networks, anticipating disruption, testing scenarios, optimizing responses and supporting recovery.

However, the study warns that digital twins do not automatically make transport systems resilient. Only 10 studies, 16.4% of the analytical corpus, reported directly verified resilience outcomes. Most demonstrated enabling functions through simulations, prototypes or conceptual frameworks rather than real-world performance during actual disruption.

From Virtual Copy to Urban Control Layer

A digital twin is often described as a virtual representation of a physical asset or system. In transport, that could mean a bridge, road corridor, railway station, port, bus network or entire mobility ecosystem. The review shows, however, that the term is frequently used too loosely.

A static digital model is not the same as a digital twin. A digital shadow automatically receives data from the physical system but does not necessarily send information or decisions back. A mature digital twin involves continuous or near-real-time synchronization and some form of bidirectional interaction. Advanced versions may support automated rerouting, adaptive traffic control, infrastructure regulation or operational reconfiguration.

This difference is not semantic. A model that visualizes traffic conditions may help planners understand congestion, but a synchronized twin can potentially detect disruption, test interventions and recommend action while events unfold.

The review identifies five dominant functions: monitoring, simulation, prediction, optimization and decision support. Sensors, connected vehicles, weather platforms, mobile applications and road infrastructure feed data into the digital layer. The twin then converts those inputs into traffic-state estimates, risk forecasts, scenario tests or operational recommendations.

In road networks, this may mean detecting abnormal congestion or evaluating alternative routes. In ports, it may support rescheduling and resource allocation under uncertainty. In rail systems, it can model passenger movements and assess the consequences of service interruption. For infrastructure managers, it can assist condition monitoring and predictive maintenance.

The strategic shift is clear. Digital twins are no longer being developed only to represent transport systems. They are increasingly positioned as cyber-physical control layers linking observation, experimentation and action.

The Resilience Case Is Stronger in Theory Than in Practice

The review separates potential resilience functions from demonstrated resilience outcomes. Monitoring can improve situational awareness. Prediction can identify weak signals. Simulation can test alternative responses. Optimization can help reallocate resources or reroute traffic. Decision-support tools can reduce uncertainty. Yet none of these functions proves that a transport network can absorb disruption, maintain essential service or recover faster.

The evidence base remains heavily weighted toward simulation. Of the 61 studies, 19 were simulation, optimization or computational studies. Fifteen were conceptual or review-based, while prototypes and field-data applications accounted for 13 studies each. Only one involved a real-world operational deployment, and even that study did not measure resilience during an actual disruption.

The authors classified 36 studies, or 59%, as providing validated evidence for enabling functions. Another 15, or 24.6%, offered conceptual or review evidence. Only 10 studies directly verified resilience-related outcomes such as service continuity, recovery, adaptive rerouting or reconfiguration.

This gap has major implications for public investment. Cities may purchase sophisticated platforms that perform well in controlled environments but struggle under real-world conditions involving incomplete data, institutional conflict, cyberattack, cascading infrastructure failure or unpredictable human behavior.

The literature is also uneven across transport sectors. Roads, highways, traffic corridors, rail systems and ports receive more attention than open multimodal networks. Bounded systems are easier to instrument and model. Citywide mobility systems are more difficult because they involve public transport agencies, private operators, emergency services, logistics firms, pedestrians and informal transport—all governed by different rules and data systems.

The result is a field rich in technological possibility but still short of comparable evidence on whether digital twins preserve throughput, shorten recovery, protect vulnerable users or maintain minimum mobility during disruption.

Resilience Depends on Institutions, Not Algorithms Alone

The review argues that digital twins should be understood as socio-technical systems. Their success depends on three groups of conditions: technical, organizational and institutional. The technical conditions are demanding. Digital twins require reliable sensors, timely data, interoperability, synchronized updates, valid models and sufficient computing capacity. They must also withstand cyber threats. Because they connect vehicles, infrastructure, communications, cloud platforms and control systems, they can expand the attack surface of the transport network.

Bad data can be as damaging as missing data. A model trained on incomplete traffic patterns may recommend the wrong diversion. A desynchronized system may present an outdated picture during an emergency. A compromised sensor or manipulated data stream could mislead operators precisely when rapid decisions are required.

Organizational capacity is equally important. Transport operators must know how to interpret digital outputs and when to override them. Agencies need clear emergency protocols, decision rights and lines of accountability. Data-sharing agreements must be established before a crisis, not negotiated during one.

Institutional coordination is the hardest layer. Smart-city transport spans multiple authorities, private operators and jurisdictions. A digital twin may identify the optimal response, but implementation can fail if agencies cannot agree, regulations prevent data exchange or procurement contracts lock cities into incompatible systems.

This is particularly relevant for developing countries and fast-growing cities in the Global South. Digital twins could help authorities anticipate flood-related road closures, manage scarce infrastructure and test investments before construction. But weak data systems, limited sensor coverage, fragmented institutions and high maintenance costs could turn ambitious platforms into expensive dashboards with little operational value.

For policymakers, the priority should therefore be use-case discipline. Cities should begin with clearly defined problems, such as flood disruption, bridge failure, major-event traffic or public transport interruption, and assess whether a digital twin improves measurable outcomes. Procurement should focus on throughput retained during disruption, travel-time degradation, recovery time, rerouting effectiveness and service continuity rather than visual sophistication.

Smart Cities Need Proof, Standards and Public Accountability

The review does not dismiss digital twins, but provides a more credible path for their development. In the short term, governments should invest in interoperable data standards, cybersecurity, model validation and operator training. Digital twins should be integrated into emergency planning, not isolated within technology departments. Transport agencies should run stress tests using realistic scenarios, including communication failure, cyberattack, extreme weather and cascading disruption.

Over the longer term, cities need governance frameworks covering privacy, data ownership, automated decision-making and liability. The growing integration of human, vehicle and traffic data raises serious questions about surveillance and exclusion. Informal mobility, low-income neighborhoods and travelers without connected devices may be poorly represented, producing technically efficient but socially distorted decisions.

The review also exposes a research agenda. Future studies should compare digital twins with simpler machine-learning, simulation or forecasting tools under identical disruption scenarios. Researchers need common resilience indicators and more longitudinal evidence across cities, transport modes and levels of technological maturity.

Greater attention must also be paid to post-disruption learning. A mature digital twin should not only help manage a crisis; it should preserve information about what happened, assess which interventions worked and improve future thresholds, models and response protocols.

By reasonable inference, the technology connects with SDG 9 on resilient infrastructure, SDG 11 on sustainable cities and SDG 13 on climate action. However, digital twins contribute to these goals only when they improve real public outcomes. Efficiency under normal conditions is not the same as resilience under stress.