As Europe prepares for a future of smarter, safer, and more sustainable transport networks, the EvoRoads project is progressing rapidly across its work packages. One of its central pillars is Work Package 2 (WP2) – Advanced infrastructure monitoring and predictive maintenance tools, responsible for equipping road operators with a new generation of cyber-physical tools for monitoring, prediction, and safety management.

WP2 represents the data-intelligence core of EvoRoads, delivering technologies that detect road defects, assess hazards, forecast infrastructure deterioration, and support timely maintenance. This intelligence directly feeds the EvoRoads platform, dashboards, and Digital Twin, enabling more proactive and efficient road operations across diverse European environments.

Deliverables Submitted So Far: Establishing the First WP2 Ecosystem

WP2 has already delivered two major milestones: D2.1 and D2.2, together forming the first operational version of the WP2 toolset.

D2.1 – Infrastructure Monitoring Tools V1

This deliverable introduces a suite of AI-powered monitoring tools capable of identifying cracks, potholes, degraded signage, pavement vibrations, deformations, and hazards captured through CCTV (Closed-Circuit Television), drones, and even accident-risk estimation. It describes how these tools function, the datasets behind them, and their level of readiness for deployment across the project’s pilot sites.

D2.2 – Predictive Maintenance & On-the-Edge Safety Tools V1

D2.2 presents the initial prototype of the project’s predictive maintenance framework, including deterioration prognostics, optimisation models, and post-repair analysis capabilities. It also introduces two edge-level safety systems already validated in real environments, providing the foundation for integration into the EvoRoads digital ecosystem.

Together, these deliverables mark a significant transition from concept to functioning technology—creating the backbone of the monitoring and predictive ecosystem covered by the work package

Unexpected Insights and Emerging Synergies

While developing these tools, the WP2 partners observed important and sometimes unexpected synergies. Combining diverse data sources—UAV (Unmanned Aerial Vehicles) imagery, satellite radar, stereo cameras (OAK-D), and connected vehicles—resulted in much higher detection accuracy than when using any single source independently.

Early tests showed that vibration data from public buses can reveal pavement issues that visual inspections often miss, highlighting strong potential for connected-vehicle sensing. In addition, although the crowdsourcing component will be launched in the next phase, experience from similar services suggests high citizen engagement, especially valuable for rural and low-traffic roads where official inspections are infrequent.

These insights are helping shape a more robust and inclusive monitoring ecosystem.

Contributing to EvoRoads’ Safety Mission

Safety is at the heart of EvoRoads, and WP2’s results already significantly advance this mission. Its tools help:

• Detect hazards earlier, from cracks and subsidence to debris, animals, and obstructions
• Predict failures before they happen, enabling timely maintenance
• Extend monitoring to remote or low-visibility areas
• Provide actionable intelligence that reduces the risk of accidents
• Deliver more efficient and resource-optimised infrastructure management

WP2’s integrated approach is turning advanced sensing and AI into real, tangible safety benefits for European citizens.

Serving All Road Types – From Cities to Rural Networks

EvoRoads is not focused on urban networks alone. Rural and secondary roads—with long distances, limited inspections, and unique hazards—are equally important. To address them, the solutions provided in WP2 form a multi-sensor, infrastructure-light monitoring ecosystem:

1. Tools requiring no fixed roadside infrastructure: UAV monitoring, InSAR, connected-vehicle vibration data, and crowdsourced reports all work effectively in remote or low-density areas where installing equipment is impractical. These tools give coverage over long distances and dispersed networks.
2. Tools built specifically for low-infrastructure conditions:
   
   
   1. Hazard-Aware Infrastructure Monitoring (HAIM) is designed for rural deployment: low-cost, solar-capable, and camera-free.
   2. Artificial Intelligence for Road Hazard Detection (AIRHD) includes synthetic training scenarios for rural hazards like fog, vegetation, and poor lighting.

3. Tools trained on rural-specific hazards: AI models detect issues common outside urban areas: vegetation overgrowth, debris, animal crossings, unstable shoulders, and subsidence.
4. Pilot alignment: Our Living labs in all sites (Spain, Italy, Romania, Latvia) contain secondary or rural segments, and every pilot already has multiple WP2 tools mapped to these needs.

What Comes Next? Advancing Toward Full Integration and Pilot Validation

Following the first round of deliverables, WP2 now enters a phase of refinement, integration, and real-world evaluation. The next steps include:

• Finalising all tools based on lessons from the first development cycle
• Integrating the solutions into the EvoRoads platform and Data Space, ensuring that each tool’s outputs flow seamlessly into the dashboards, decision-support modules, and the Digital Twin.
• Deploying and testing the full toolset across the four Living Labs
• Strengthening multi-source data fusion for more reliable monitoring
• Enhancing multi-source data fusion, combining UAV imagery, connected-vehicle data, InSAR deformation signals, crowdsourced reports, and edge-based sensing to achieve more reliable and comprehensive monitoring.
• Evaluating and improving edge devices (HAIM and AIRHD) in operational environments

These activities build directly on the foundations established in D2.1 and D2.2 and will lead to a complete, validated release of the WP2 technologies during the project’s pilot demonstrations.