The researchers focus on Cyber Physical Human Systems, or CPHS—an extension of traditional cyber-physical systems that places humans in the loop as active participants. These systems, deployed across smart cities to manage water, power, transportation, and healthcare, typically span an edge-fog-cloud computing continuum with multiple abstraction levels covering services, applications, physical infrastructure, and the sense-compute-actuate information layer. Despite providing some adaptivity to changing conditions, current CPHS are normally purpose-built, designed to operate in their own application context and to react only to changes happening in that context. This makes them brittle in the face of extreme events, since individual systems cannot seamlessly interoperate with adjacent ones to support adaptive responses or joint deployment during emergencies.

By repurposing, the authors mean reusing the exact same infrastructure for applications and scenarios that individual CPHS were not originally designed for. They define infrastructure resilience as the capacity of community-scale systems to absorb, recover from, and adapt to disruptive events—natural disasters, infrastructure failures, or unexpected operational demands—while continuing to deliver essential services. The team acknowledges that cross-integrating independently designed systems may introduce new inter-dependencies that could lead to cascading disruptions if not carefully managed, but argues that the benefits of broader situational awareness, resource sharing, and coordinated response outweigh these risks when supported by appropriate orchestration, governance, and modularity mechanisms that contain failure propagation across system boundaries.

To ground the vision, the researchers draw on two case studies informed by over a decade of projects executed on both sides of the Atlantic. The first is CareDEX, a US NSF project focused on information exchange for healthcare resilience. Older adults suffer a disproportionate level of fatalities and injuries during emergencies and natural disasters, often unable to shelter safely in place or evacuate on their own due to physical conditions such as the need for life-sustaining equipment or impaired mobility, and cognitive afflictions like dementia or Alzheimer’s disease. CareDEX was designed to create information preparedness for aging communities by enabling rapid and secure exchange of personalized care information between first responders, caregivers in senior housing facilities, and older adults themselves.

The CareDEX platform includes decision support tools built atop the TIPPERS smart-space IoT data management middleware, creating a digital twin that captures spatial, temporal, and human aspects of the physical environment. Caregivers gain access to personalized information about residents’ care needs, health conditions, and facility resources, while a Regional Impact Module provides interactive visualization of disaster impacts at regional scale, drawing on public hazard data sources including the HAZUS toolkit from USGS, wildfire information from the National Interagency Fire Center, and NOAA’s Hazard Mapping System Fire and Smoke Product. The system balances individual privacy with authorized release of information to responders when needed, ensuring that sensitive health profiles and facility details are protected while still empowering effective triage, shelter-in-place decisions, and safe evacuations.

The second case study comes from the EU project MYRTUS, which demonstrates end-to-end orchestration across the cloud-to-edge continuum for monitoring, analysis, and control of urban mobility. Deployed at a traffic-intensive urban intersection in the city of Helmond, the system equips traffic lights with smart cameras, edge computing devices, and communication equipment to detect and track vehicles and vulnerable road users such as pedestrians and cyclists, estimating collision risks and issuing warnings to approaching vehicles. All sensors and actuators connect to an embedded Jetson Orin device installed in the roadside traffic control cabinet, operating as a fog server, with secure VPN connections providing access to cloud resources. The project considers three operational modes: road user detection performed entirely within cameras, partial processing delegated to the fog server including fusion of detections from multiple cameras and privacy-preserving detection based on blurred video streams, and full cloud processing of blurred streams to derive long-term traffic statistics.

When disasters strike, both systems could be repurposed in ways their designers never envisioned. CareDEX, originally built for safe evacuation of seniors, could coordinate multiple site evacuations across schools, hospitals, offices, concert halls, and stadiums during a large-scale disaster, matching populations with specific needs to appropriate destination sites. Meanwhile, the urban traffic monitoring infrastructure from MYRTUS could be repurposed to detect traffic signal and sensor failures, assess road accessibility and conditions, and identify their impact on transportation corridors. Conjoined together, the repurposed systems enable safe and efficient city-wide evacuation plans: the status of road infrastructure informs the secure information exchange system, while congestion caused by evacuating populations requires reconfiguration of traffic infrastructure to help ingress and egress of emergency vehicles.

The researchers identify four key enablers required to make repurposing work at scale. Lightweight containerization technologies must explicitly target edge and embedded nodes with limited resources and heterogeneous computing nodes, including reconfigurable hardware such as FPGAs. Seamless deployment mechanisms must operate consistently across edge, fog, and cloud layers and across heterogeneous operating systems and middleware stacks. Intelligent orchestration engines must automate discovery and placement of software components across layers while complying with resource, security, and communication constraints—MYRTUS is developing a cognitive orchestration engine that uses machine learning techniques and distributed optimization to explore the design space of feasible configurations. Finally, security and trust primitives must remain effective when infrastructures are repurposed, federated, or shared across organizational boundaries, with dynamic access control policies that grant temporary access to first responders only as long as needed and revoke it afterward, ensuring no private data persists after the evacuation process.

Despite the promise, the current state of these enablers reveals significant gaps. Cloud platforms provide mature support for containerization, deployment automation, and orchestration, but edge and fog layers—precisely where repurposing is often most valuable due to their proximity to the physical environment—remain limited and fragmented. Vendor lock-in manifests at multiple levels, from hardware accelerators with proprietary toolchains to platform-specific APIs and closed monitoring solutions, all of which constrain the ability to move software components between execution environments. The researchers also highlight a third dimension of interoperability beyond horizontal and vertical: organizational interoperability, which supports federation across infrastructures owned and operated by different stakeholders and is fundamental for disaster response scenarios. Initiatives such as the EU-CEI and DISCOVER-US are taking concrete steps to identify and tackle these open challenges, investing in standardization promotion and transatlantic collaboration respectively.

Looking ahead, the researchers see artificial intelligence as a promising direction for extracting live situational awareness and long-term impacts. Federated learning approaches like FLIPS manage data and participant heterogeneity by clustering devices based on their data distributions, allowing heart rhythm and fall detection models to be trained without compromising sensitive personal data. Newer streaming federated learning systems like ShiftEx use a Mixture of Experts approach to handle continually evolving models and non-stationary inputs as patient data evolves over time. The team also points to the need for aggressive optimization of privacy-preserving technologies such as homomorphic encryption and secure multi-party computation to cope with changing environments, and for middleware approaches like BATCHIT that leverage intelligent batching to enhance streaming performance in resource-constrained IoT smart spaces. Ultimately, the authors note that human and social factors—including cognitive biases in decision-making within interdependent infrastructures—can significantly affect outcomes, and that modularity as a complementary strategy to cross-integration represents a promising direction for future research aimed at containing cascading effects of erroneous decisions while still capturing the benefits of coordination.

Subject of Research: A vision article proposes repurposing siloed smart city computing infrastructures to improve community-scale resilience during disasters.

Article Title: Towards community-scale resilience via repurposable computing continuum infrastructures

Article References: Towards community-scale resilience via repurposable computing continuum infrastructures. (n.d.). https://doi.org/10.3389/fcpxs.2026.1812652

Image Credits: AI Generated

DOI: 10.3389/fcpxs.2026.1812652

Keywords: computing continuum, cyber-physical-human systems, smart cities, disaster resilience, edge computing, fog computing, CareDEX, MYRTUS, federated learning, privacy, orchestration, containerization