D2.3 – Technical Requirements and Digital Ecosystem Architecture

@iDriving Consortium 2024 – 2027 – https//idriving-project.eu Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies D2.3 – Technical Requirements and Digital Ecosystem Architecture Author(s): Ioannis Tampakis, George Lalas, Dariya Rublova Leading Partner: Netcompany S.A. Version - Status: V0.9 Submission Date: 06/2025 Dissemination Level: PUBLIC Disclaimer: Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or CINEA. Neither the European Union nor CINEA can be held responsible for them. Copyright message: ©iDriving Consortium, 2024-2027. This deliverable contains original unpublished work except where clearly indicated otherwise. Acknowledgement of previously published material and of the work of others has been made through appropriate citation, quotation, or both. Reproduction is authorised provided the source is acknowledged. @iDriving Consortium – 101147004 Page 2 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Report information Purpose of the Report This report outlines the technical requirements and system architecture for iDriving, translating user needs into technical specifications. It provides a foundation for the development, integration, and validation of all project components. Relevant Work package: WP2 Relevant Task: T2.5 Technical requirements specifications and ecosystem architecture Contributors: INTRA, CERTH, TEKNIKER, DREVEN, MBL, TUC, UNI.EIFFEL, INFRA PLAN, AUSTRIATECH, ACCELI, SIMAVI Next version Title: NA Next version Date: NA Official Submission Date: 30/06/2025 Actual Submission Date: 30/06/2025 ABOUT iDRIVING iDriving, a 3-year Horizon-funded project, unites 17 European partners in a mission to enhance road safety. The project focuses on transforming urban and secondary rural road infrastructure through innovation. It aligns with the EU’s goals for smart transport and actively embraces emerging technologies. Key areas of impact include enhancing driver behaviour, improving infrastructure safety, and empowering first responders. Through a comprehensive Safety Criteria Catalogue, innovative sensors, AI-based warnings, and a Digital Twin, iDriving paves the way for safer roads. The iDriving consortium consists of the following partners: No Participant organisation name Short name Country 1 ETHNIKO KENTRO EREVNAS KAI TECHNOLOGIKIS ANAPTYXIS CERTH EL 2 POLYTECHNEIO KRITIS TUC EL 3 AUSTRIATECH - GESELLSCHAFT DES BUNDES FUR AT TECHNOLOGIEPOLITISCHE MASSNAHMEN GMBH AUSTRIATECH AT 4 UNIVERSITE GUSTAVE EIFFEL UNI.EIFFEL FR 5 FUNDACION TEKNIKER TEKNIKER ES 6 INGARTEK CONSULTING SL ING ES 7 INFRA PLAN KONZALTNIG JDOO ZA USLUGE INFRA PLAN HR 8 ACCELIGENCE LTD ACCELI CY @iDriving Consortium – 101147004 Page 3 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies 9 NETCOMPANY S.A. INTRA LU 10 SOFTWARE IMAGINATION & VISION SRL SIMAVI RO 11 ALP.Lab GmbH ALP.LAB AT 12 MUNICIPALITY OF ALBA IULIA AIM RO 13 DRAXIS RESEARCH VENTURES ASTIKI MI EL KERDOSKOPIKI ETAIRIA DREVEN EL 14 PRAVO I INTERNET FOUNDATION LIF BG 15 DIMOS THESSALONIKIS THESSALONIKI EL 16 GRAD KARLOVAC COK HR 17 MOBILYSIS SARL MOBILYSIS SARL CH Version History Version Date Author Partner Description 0.1 25/11/2024 Ioannis Tampakis INTRA First version of ToC 0.2 18/03/2025 Ioannis Tampakis, George Lalas INTRA Updated ToC 0.3 19/05/2025 Dimitris Perikleous, Andreas Lefkatis (ACCELI), Alexandros Petropoulos, Stavros Paspalakis, Panos Vrachnos, Emmanuel Raptis, Konstantinos Ioannidis, Alexandros Sfyridis, Nikos Dourvas (CERTH), Mostafa Ameli, Thomas Bapaume, Latifa Oukhellou (UNI.EIFFEL), Emmanouil Barmpounakis, Dimitrios Tsitsokas, Jasso Espadaler Clapes (MBL), Panagiotis Tsalis (Thessaloniki) Susana Ferreiro, Gonzalo Gil, Iker Narbaiza, Elena García (TEK) ACCELI, CERTH, UNI.EIFFEL, MBL, THESSALONIKI, TEKNIKER, TUC, SIMAVI, DREVEN, INFRA PLAN, ALP.LAB, AUSTRIATECH All relevant inputs from WP2, WP3, WP4 and WP5 Tasks @iDriving Consortium – 101147004 Page 4 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Diana Gornea, Daniel Gherghiceanu, Lavinia Popa (SIMAVI), Vasileios Markantonakis (TUC), Zoi Dimitriadou (DREVEN), Irina Stipanovic (INFRA PLAN), Mohamed Berrazouane (ALP.LAB), Helena Korndoerfer (AUSTRIATECH) 0.4 20/05/2025 Ioannis Tampakis INTRA iDriving Architectural views added 0.5 28/05/2025 Ioannis Tampakis, George Lalas INTRA Updated introduction, methodology, conclusion, executive summary, and references. 0.6 05/06/2025 Relevant partners Relevant partners Further updates from the contributing partners 0.7 24/06/2025 Dimitrios Tsitsokas (MBL), Irina Stipanovic (INFRA PLAN) MBL, INFRA PLAN Internal review 0.8 26/06/2025 Relevant partners Relevant partners All comments addressed 0.9 26/06/2025 Ioannis Tampakis, Dariya Rublova (INTRA) INTRA Pre-final document sent to Coordinator 1.0 30/06/2025 Alexandros Sfyridis CERTH Submission to EC @iDriving Consortium – 101147004 Page 5 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Table of Contents 1 Introduction ............................................................................................................ 13 1.1 Deliverable context .................................................................................................. 13 1.2 Approach to defining the requirements .......................................................... 14 2 Overview of the iDriving system ..................................................................... 15 2.1 System context and purpose ................................................................................ 15 2.2 System Description and Main Capabilities ...................................................... 15 2.3 Summary of use cases, scenarios, and user requirements ........................16 2.3.1 Use Case 1.1: Graz, Austria ......................................................................................................................... 16 2.3.2 Use Case 1.2: Nevers, France .............................................................................................................. 17 2.3.3 Use Case 2.1: Karlovac, Croatia ......................................................................................................... 18 2.3.4 Use Case 2.2: Thessaloniki, Greece ............................................................................................... 18 2.3.5 Use Case 3.1: Alba Iulia, Romania .................................................................................................... 19 2.3.6 Use Case 3.2: Bizkaia, Spain .............................................................................................................. 20 3 Technical requirements ...................................................................................... 21 3.1 Functional & Non-Functional requirements .................................................... 21 3.1.1 General Non-Functional requirements .......................................................................................... 21 3.1.2 Requirements related to WP4............................................................................................................. 35 3.1.3 Requirements related to WP5 ............................................................................................................ 50 4 Technical & user requirements Mappings .................................................. 59 5 Overall Architecture & Components ............................................................. 67 5.1 Methodology for Defining the Architecture ................................................... 67 5.2 iDriving Context View ............................................................................................. 67 5.3 iDriving Containers View ...................................................................................... 68 5.3.1 Containers View – Consolidated ........................................................................................................ 69 5.3.2 Containers View Per Use Case ......................................................................................................... 71 5.3.2.1 Use case 1.1 Graz, Austria ...................................................................................................... 71 5.3.2.2 Use Case 1.2 Nevers, France................................................................................................... 72 5.3.2.3 Use Case 2.1 Karlovac, Croatia ............................................................................................... 73 5.3.2.4 Use case 2.2 Thessaloniki, Greece .......................................................................................... 74 5.3.2.5 Use case 3.1 Alba Iulia, Romania ............................................................................................ 75 @iDriving Consortium – 101147004 Page 6 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies 5.3.2.6 Use case 3.2 Bizkaia, Spain ..................................................................................................... 76 5.4 iDriving Components Views & Specifications ................................................ 76 5.5 iDRIVING Process Views ...................................................................................... 128 5.5.1 Tekniker Dataspace Connector......................................................................................................... 129 5.5.2 Route Guidance tool ........................................................................................................................... 130 5.5.3 Signal Control Tool ................................................................................................................................ 131 5.5.4 Mobile Application & In-Vehicle Application ....................................................................... 132 5.5.5 Aerial Surveillance UAV System ................................................................................................... 133 5.5.6 Autonomous UAV Deployment for Area Coverage ........................................................ 134 5.5.7 Seat Belt and Cell Phone Detection Tool............................................................................... 135 5.5.8 License Plate Detection Tool ......................................................................................................... 136 5.5.9 Helmet Detection Tool....................................................................................................................... 137 5.5.10 Zebra Crossing Detection Tool ..................................................................................................... 138 5.5.11 Fallen Tree and Rockslide Detection Tool ............................................................................. 138 5.5.12 Crashed Vehicle Detection Tool .................................................................................................. 139 5.5.13 Red Light Violation Detector ........................................................................................................ 140 5.5.14 Improper Lane Usage Detector .................................................................................................... 141 5.5.15 Zebra Crossing Violation Detector ............................................................................................. 142 5.5.16 Abrupt Movement Detector .......................................................................................................... 143 5.5.17 Pothole Detection and Severity Assessment Tool .......................................................... 144 5.5.18 Weather Prediction Tool (WRF and Data Assimilation) ............................................... 145 5.5.19 AI-Based Real-Time Weather Alert System .........................................................................146 5.5.20 3D-SMART Tool ....................................................................................................................................... 147 5.5.21 ClaireSITI Platform ................................................................................................................................ 147 5.5.22 SUMO (Simulation of Urban Mobility) ..................................................................................... 148 5.5.23 CARLA (Autonomous Driving Simulation Platform) ......................................................149 5.5.24 AI-Optimized Maintenance through Digital Twin ........................................................... 150 5.5.25 Digital Twin-Based Control Centre with XR features for Enhanced Situational Awareness ...................................................................................................................................................................... 155 6 Conclusion / Future Work ............................................................................... 156 References ................................................................................................................... 157 7 PROJECT FACTS .................................................................................................. 158 @iDriving Consortium – 101147004 Page 7 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Table of tables Table 1 - Use Case 1.1: Graz, Austria ............................................................................................................................... 16 Table 2 - Use Case 1.2: Nevers, France ........................................................................................................................ 17 Table 3 - Use Case 2.1: Karlovac, Croatia ................................................................................................................... 18 Table 4 - Use Case 2.2: Thessaloniki, Greece ......................................................................................................... 18 Table 5 - Use Case 3.1: Alba Iulia, Romania .............................................................................................................. 19 Table 6 - Use Case 2.3: Bizkaia, Spain ....................................................................................................................... 20 Table 7 – Technical requirement to user requirement mapping ........................................................... 59 Table 8 - User requirement to technical requirement mapping ........................................................... 62 Table 9 - Tekniker Dataspace Connector - Architectural Design and Description .................... 78 Table 10 - Route Guidance tool Architectural Design and Description ..............................................79 Table 11 - Signal Control Tool – Architectural Design and Description ............................................... 82 Table 12 - Mobile Application – Architectural Design and Description ...............................................85 Table 13 - In-Vehicle Application – Architectural Design and Description ...................................... 86 Table 14 - Aerial Surveillance UAV System – Architectural Design and Description ................. 89 Table 15 - Autonomous UAV Deployment for Area Coverage – Architectural Design and Description ........................................................................................................................................................................ 91 Table 16 - Seat Belt and Cell Phone Detection Tool – Architectural Design and Description ................................................................................................................................................................................................. 94 Table 17 - Count Cars Detection Tool – Architectural Design and Description ............................. 95 Table 18 - License Plate Detection Tool – Architectural Design and Description ....................... 96 Table 19 - Helmet Detection Tool – Architectural Design and Description ..................................... 98 Table 20 - Zebra Crossing Detection Tool – Architectural Design and Description .................. 99 Table 21 - Fallen Tree and Rockslide Detection Tool – Architectural Design and Description ............................................................................................................................................................................................... 100 Table 22 - Crashed Vehicle Detection Tool – Architectural Design and Description ............... 102 Table 23 - Red Light Violation Detector – Architectural Design and Description ..................... 103 Table 24 - Improper Lane Usage Detector – Architectural Design and Description .............. 104 Table 25 - Zebra Crossing Violation Detector – Architectural Design and Description ......... 105 Table 26 - Abrupt Movements Detector – Architectural Design and Description .................... 107 Table 27 - Pothole Detection and Severity Assessment Tool – Architectural Design and Description .....................................................................................................................................................................109 @iDriving Consortium – 101147004 Page 8 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Table 28 - Weather Prediction Tool (WRF and Data Assimilation) – Architectural Design and Description ...................................................................................................................................................................... 110 Table 29 - AI-Based Real-Time Weather Alert System – Architectural Design and Description ................................................................................................................................................................................................. 113 Table 30 - 3D-Smart Tool – Architectural Design and Description ....................................................... 115 Table 31 - Dynamic Monitoring Platform – Architectural Design and Description .................... 117 Table 32 - SUMO: Digital Twin Powered Predictive Safety Measures and Warning Systems – Architectural Design and Description .......................................................................................................... 118 Table 33 - CARLA: Digital Twin Powered Predictive Safety Measures and Warning Systems – Architectural Design and Description ......................................................................................................... 120 Table 34 - AI-Optimized Maintenance Through Digital Twin – Architectural Design and Description ..................................................................................................................................................................... 123 Table 35 - Dynamic Risk Assessment Module – Architectural Design and Description ........ 124 Table 36 - Health and Logistics Management Module – Architectural Design and Description ................................................................................................................................................................................................ 125 Table 37 - Maintenance Scheduling Tool – Architectural Design and Description .................. 127 Table 38 - Digital Twin-Based Control Centre – Architectural Design and Description ........ 127 Table Of Figures Figure 1 - iDriving System Overview ........................................................................................................................... 16 Figure 2 - iDriving Context View ................................................................................................................................. 68 Figure 3 - Consolidated iDriving Containers view ........................................................................................... 70 Figure 4 - Containers View Colour Legend .......................................................................................................... 70 Figure 5 - Use Case 1.1 Containers View .................................................................................................................... 71 Figure 6 - Use Case 1.2 Containers View .................................................................................................................. 72 Figure 7 - Use Case 2.1 Containers View .................................................................................................................. 73 Figure 8 - Use Case 2.2 Containers View ................................................................................................................ 74 Figure 9 - Use Case 3.1 Containers View .................................................................................................................. 75 Figure 10 - Use Case 3.2 Containers View ...............................................................................................................76 Figure 11 - Tekniker dataspace connector sequence diagram ............................................................... 129 Figure 12 - Route Guidance tool sequence diagram .................................................................................... 130 Figure 13 - Signal Control tool sequence diagram .......................................................................................... 131 Figure 14 - Mobile Application & In-Vehicle Application sequence diagram ............................... 132 Figure 15 - Aerial Surveillance UAV System sequence diagram ............................................................ 133 @iDriving Consortium – 101147004 Page 9 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Figure 16 - Autonomous UAV Deployment for Area Coverage sequence diagram ................. 134 Figure 17 - Seat Belt and Cell Phone Detection Tool sequence diagram ....................................... 135 Figure 18 - License Plate Detection Tool sequence diagram .................................................................. 136 Figure 19 - Helmet Detection Tool sequence diagram ............................................................................... 137 Figure 20 - Zebra Crossing Detection Tool sequence diagram ............................................................ 138 Figure 21 - Fallen Tree and Rockslide Detection Tool sequence diagram ...................................... 138 Figure 22 - Crashed Vehicle Detection Tool sequence diagram .......................................................... 139 Figure 23 - Red Light Violation Detector sequence diagram ................................................................ 140 Figure 24 - Improper Lane Usage Detector sequence diagram ........................................................... 141 Figure 25 - Zebra Crossing Violation Detector sequence diagram ..................................................... 142 Figure 26 - Abrupt Movement Detector sequence diagram .................................................................. 143 Figure 27 - Pothole Detection and Severity Assessment Tool sequence diagram .................. 144 Figure 28 - Weather Prediction Tool sequence diagram .......................................................................... 145 Figure 29 - AI-Based Real-Time Weather Alert System sequence diagram .................................146 Figure 30 - 3D-SMART Tool sequence diagram ............................................................................................... 147 Figure 31 - ClaireSITI Platform sequence diagram ......................................................................................... 147 Figure 32 - SUMO sequence diagram ................................................................................................................... 148 Figure 33 - CARLA sequence diagram ...................................................................................................................149 Figure 34 - AI-Optimized Maintenance through Digital Twin - Global Component sequence diagram ............................................................................................................................................................................ 150 Figure 35 - AI-Optimized Maintenance through Digital Twin – Module 1.1 Risk Assessment sequence diagram ..................................................................................................................................................... 151 Figure 36 - AI-Optimized Maintenance through Digital Twin – Module 1.2 Health & Logistics Management sequence diagram ................................................................................................................... 153 Figure 37 - AI-Optimized Maintenance through Digital Twin - Module 1.3 Maintenance Scheduling sequence diagram ........................................................................................................................ 154 Figure 38 - Digital Twin-Based Control Centre with XR features for Enhanced Situational Awareness sequence diagram ......................................................................................................................... 155 @iDriving Consortium – 101147004 Page 16 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Figure 1 - iDriving System Overview This digital ecosystem enables seamless information flow among system components and stakeholders, supporting both centralized control and decentralized, on-site action. 2.3 Summary of use cases, scenarios, and user requirements The following sub-chapters provide an overview of the six use cases, including scenario descriptions and user requirements. 2.3.1 Use Case 1.1: Graz, Austria Table 1 - Use Case 1.1: Graz, Austria Scenario Title Description Congestion Prediction and Dynamic Rerouting Continuous monitoring of traffic flow is conducted using various sensors and AIenabled cameras. Predictive models analyse real-time and historical data to forecast congestion. Based on these predictions, alternative routes are communicated to road users via VMS and mobile or In-vehicle applications to optimize traffic distribution. @iDriving Consortium – 101147004 Page 17 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Automated Detection of Road User Violations Video analytics systems automatically detect violations such as non-usage of helmets or seatbelts, traffic light infractions, and unauthorized lane usage, including wrong way driving or improper use of turning lanes. Identified violations are documented for road operators and communicated to relevant road users in order to increase compliance. Safety-Critical Incidents Detection The system continuously analyses sensors and video data to identify dangerous or critical situations. Upon detection, immediate alerts are issued to road users and road operators to enable prompt intervention and minimize potential risks. 2.3.2 Use Case 1.2: Nevers, France Table 2 - Use Case 1.2: Nevers, France Scenario title Description Deployment of Monitoring systems Install cameras at key locations like intersections, parking zones, and zebra crossings, while deploying UAVs over critical areas, including bicycle lanes and pedestrianheavy zones, to provide a comprehensive ground and aerial view of roads Real Time integration and analysis Integrate ground camera and UAV data into iDriving’s Digital Twin, using AI to analyse traffic patterns, detect infractions, identify risky behaviours, and highlight areas for infrastructure improvements. Communication and alerts to road users Activate messaging on road or online systems to alert road users of potential incidents. Notify drivers approaching zebra crossings to yield via VMS or in-vehicle messages. Use VLC for communication and warn distracted cyclists to enhance compliance and safety. @iDriving Consortium – 101147004 Page 18 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies 2.3.3 Use Case 2.1: Karlovac, Croatia Table 3 - Use Case 2.1: Karlovac, Croatia Scenario title Description Detection of road damages (potholes, cracks) Deploy several technologies to support inspection procedure: data collection using cameras on inspection vehicles, fixed cameras on critical sections, like intersections or sections with heavy traffic, and UAVs over remote areas. Environmental monitoring using smart weather stations, to correlate road damages occurrence. Provide maintenance plan to infrastructure managers for repair of detected potholes Develop DT-based optimized model for maintenance plan while considering traffic flow (cameras and traffic counters data) and weather forecast (weather station) Communication and alerts to road users before and during maintenance works Activate messaging on road or online systems to alert road users of maintenance road activities. Notify drivers approaching road work zones to yield via VMS or in-vehicle messages. Use VLC for communication and warn distracted cyclists to enhance compliance and safety. Monitor workers safety during maintenance works Deploy UAV during the maintenance works to detect any safety violations in the work zone areas 2.3.4 Use Case 2.2: Thessaloniki, Greece Table 4 - Use Case 2.2: Thessaloniki, Greece Scenario title Description Accumulating water from rainfall/ Gusty winds Advanced weather stations monitor weather data. Data analysed to identify patterns indicating potential road hazards, such as water accumulation that could lead to slippery conditions or flooding or strong winds for two wheelers. In-vehicle warning application, digital road signs, and the dedicated iDriving mobile application all start issuing alerts and inform on alternative routes. Traffic manager is informed Fallen Trees/Rockslides UAVs fly autonomously patrolling the risk areas to detect obstacles on the road like fallen trees or rocks. Existence of obstacles on the road @iDriving Consortium – 101147004 Page 19 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies leads to warning to iDriving platform. Warning applications and digital road signs all start issuing alerts and inform on alternative routes. Traffic manager is informed 2.3.5 Use Case 3.1: Alba Iulia, Romania Table 5 - Use Case 3.1: Alba Iulia, Romania Scenario title Description Aggressive Driving Detection Detection of aggressive behaviour in traffic (e.g., sudden lane changes, rapid acceleration/deceleration). The scenario implies using video cameras with video analytics to identify risky behaviours, (e.g., traffic sensors to detect sudden lane changes and hard braking,) dedicated Traffic Management Centre platform for analysing and processing alerts. Speeding detection Real-time speeding detection using smart cameras and speed cameras The scenario implies using video cameras for vehicle identification, smart radars for speed measurement, platform dedicated to the existing Traffic Management System in Alba Iulia, mobile application for near real-time user notification. The scenario targets speed over the legal limit detection, automatically generating alerts for drivers and sending them via the mobile app, analysis of collected data to identify high-risk areas, implementation of proactive measures to optimize traffic flow (traffic light adjustment) Speeding and Aggressive Driving Detection Combining detection of excessive speed and aggressive behaviour to provide a complete traffic risk monitoring solution. The scenario implies using video cameras and smart radars for speed and aggressive behaviour detection, traffic sensors to monitor sudden lane changes and dangerous braking, dedicated platform for traffic management system in Alba @iDriving Consortium – 101147004 Page 20 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Iulia, mobile application for near real-time alerts and notifications. The scenario targets simultaneous detection of excessive speeding and aggressive behaviour, driver notification via VMS panels and mobile app. Creating heat maps to identify high-risk areas. 2.3.6 Use Case 3.2: Bizkaia, Spain Table 6 - Use Case 2.3: Bizkaia, Spain Scenario title Description Real-time detection iDriving sensors installed in one of the vehicles involved in the accident detect and report the sudden stop. Other drivers in the area, using the iDriving in-vehicle or the mobile app, confirm the accident. Deployment of Monitoring systems Cameras installed, and UAVs dispatched over the area provide a comprehensive view of the situation, covering both ground and aerial perspectives. Digital twin generation NeRFs and visual analysis create an accurate 3D model of the zone. This real-time data is shared with local authorities, providing a virtual replication. Rerouting and planning Using historical traffic data and real-time input, iDriving’s algorithms determine the fastest route for emergency vehicles and the alternative routes for approaching drivers. Real-time alerts Through the mobile app, users are informed about the alternative routes, while VMS suggest slowing down within the area. @iDriving Consortium – 101147004 Page 21 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies 3 Technical requirements This section presents the iDriving technical requirements, grouped according to their scope and relationship to the project’s work structure. Unlike a strict split between functional and non-functional requirements, our approach is as follows: • General Non-Functional Requirements are listed in a dedicated section. These capture cross-cutting quality attributes that apply to the system, such as real-time performance, interoperability, access control, usability, data protection, and compliance. These are identified with the format TR-NFUNGEN-y, where y is a running index for general non-functional requirements. • Work Package-Specific Requirements are then detailed per Work Package (WP). For each WP, we present both functional and non-functional requirements together, reflecting the practical integration of quality attributes and system functionalities in the project’s implementation. Requirements here are labelled as follows: o TR-FUN-x-y for functional requirements, where x is the WP number and y is the index within that WP. o TR-NFUN-x-y for non-functional requirements specific to a WP, following the same numbering logic. This structure ensures that both types of requirements, functional (what the system does) and non-functional (how the system performs or must behave), are addressed in direct connection to the responsible project areas and tools. It also makes clear which requirements apply project-wide and which are WP-specific. The set of requirements can evolve during the project, due to e.g. an update of the user requirements. This may lead to additional requirements, which can be assigned consecutive higher numbers or already existing numbers with lowercase letters appended (e.g. …-1a), or to some requirements being dropped, without renumbering of the full set. 3.1 Functional & Non-Functional requirements 3.1.1 General Non-Functional requirements ID TR-NFUN-GEN-1 Name (Optional) Real-Time Performance Description Critical alerts, safety processing, and control functions must operate with real-time or near real-time latency Category Performance @iDriving Consortium – 101147004 Page 22 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Related Tasks T3.2, T3.3, T4.1, T4.2, T5.5 Related Tool - Related Use Cases All MoSCoW Scale must-have Dependencies - Input Needed - Verification Method Performance testing, simulation Comments - ID TR-NFUN-GEN-2 Name (Optional) Interoperability via Standard Formats Description All tools must exchange data using open, well-defined formats (JSON, CSV, etc.) with schemas. Category Interoperability Related Tasks All Related Tool - Related Use Cases All MoSCoW Scale must-have Dependencies - Input Needed - Verification Method Integration testing, schema validation Comments - ID TR-NFUN-GEN-3 Name (Optional) Access Control Description The platform must provide role-based access to functions (e.g., admin/operator), with clear separation of privileges. Category Security, Design Related Tasks All Related Tool - Related Use Cases All MoSCoW Scale must-have @iDriving Consortium – 101147004 Page 23 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Dependencies - Input Needed - Verification Method Access control testing, role simulation Comments E.g. Admin, operator, and user roles. ID TR-NFUN-GEN-4 Name (Optional) Usability for Critical UIs Description Mobile, infotainment, and operator UIs must comply with usability standards (e.g., non-distraction, ISO 15005), and be tested with end users. Category Usability, Design Related Tasks T3.3, T5.5 Related Tool - Related Use Cases All MoSCoW Scale must-have Dependencies - Input Needed - Verification Method Usability testing, surveys, Comments - ID TR-NFUN-GEN-5 Name (Optional) Input Data Quality Description All image/video inputs must have ≥1080p resolution and ≥5fps frame rate for accurate detection. Category Data Quality Related Tasks T3.4, T4.1 Related Tool - Related Use Cases All MoSCoW Scale must-have Dependencies - Input Needed - Verification Method Data validation @iDriving Consortium – 101147004 Page 24 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Comments - ID TR-NFUN-GEN-6 Name (Optional) Hardware Compatibility Description All edge processing modules must be deployable on an embedded device such as a NVIDIA Jetson Category Compatibility Related Tasks T4.1 Related Tool - Related Use Cases All MoSCoW Scale must-have Dependencies - Input Needed - Verification Method Deployment testing on Jetson Comments Required for field deployment ID TR-NFUN-GEN-7 Name (Optional) Network Connectivity Description The system must maintain an active internet connection to ensure real-time processing and communication. Category Availability Related Tasks All Related Tool - Related Use Cases All MoSCoW Scale must-have Dependencies - Input Needed - Verification Method - Comments Applies to UAVs, edge devices, server comms. ID TR-NFUN-GEN-8 Name (Optional) Camera Field of View @iDriving Consortium – 101147004 Page 25 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Description Cameras must be installed to maximize field of view and optimize for the detection algorithm’s needs. Category Deployment, Quality Related Tasks T4.1, T4.2 Related Tool - Related Use Cases All MoSCoW Scale should-have Dependencies - Input Needed - Verification Method Field validation, sample checks Comments - ID TR-NFUN-GEN-9 Name (Optional) Data Protection and Ethical Compliance Description The system shall handle all detection operations and data processing in accordance with GDPR requirements and established AI ethics principles (e.g., fairness, transparency, accountability). Category Security, Privacy, Ethics Related Tasks All Related Tool All Related Use Cases All MoSCoW Scale must-have Dependencies Applicable GDPR regulations; AI Ethics Frameworks (e.g., EU AI Act) Input Needed - Verification Method - Comments This NFR ensures that any collection, processing, or analysis of personal data respects user privacy rights and that AI-driven operations remain fair, transparent, and accountable. iDriving technical requirement specification ID TR-FUN-3.1-1 Name (Optional) Data Catalog Publication, Policy Negotiation, Data Access @iDriving Consortium – 101147004 Page 32 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Input Needed For real time communication and live data transmission network is required. Verification Method - Comments ACCELI can provide a 4G stick that can be connected to the ground station or directly to the drone, providing the network required for transmitting data to the cloud, if needed. This requirement is a task-specific extension of TR-NFUN-GEN-7.” ID TR-NFUN-3.4.5 Name (Optional) Compliance with Regulations Description The UAV must comply with EASA regulations in the Open Category A1/A3, namely: • Maximum Take-off Mass (MTOM): Drones in this subcategory can have a maximum weight of 25 kilograms. • No Fly Zones: Flying is prohibited over open-air assemblies of people (e.g., concerts, festivals) and densely populated areas with more than 300 people per square meter. • Visual Line of Sight (VLOS): The pilot must always maintain visual contact with the drone during the flight. • Maximum Altitude: Drones cannot fly above 120 meters above ground level (AGL) unless a special authorization is obtained. Category Non-Functional, Regulatory/compliance Related Tasks T3.4. Aerial Surveillance in Incident Management and Maintenance Tasks Related Tool iDriving Mission UAV Related Use Cases Spain, Austria, Greece, Croatia, France MoSCoW Scale Must Have Dependencies - Input Needed All required flight authorizations must be provided by the end users to the ACCELI team to conduct the missions over their designated areas. Verification Method - Comments - @iDriving Consortium – 101147004 Page 33 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies ID TR-FUN-3.5-1 Name (Optional) Mission Area Definition Description The service shall accept as input a geographical area defined as a polygon in WGS84 coordinates, including information about possible no-fly zones and/or obstacles. This input will define the spatial boundaries and constraints of the mission area, which will be used to inform and constrain UAV path planning and mission execution. Category Functional, Design, Regulatory/Compliance Related Tasks T3.5. Dynamic Resource Allocation for Optimized Safety Surveillance Related Tool UAV-based Path Planning for Coverage Operations Related Use Cases UC 1.2 Nevers, France/ UC 2.1 Karlovac, Croatia/ UC 2.2 Thessaloniki, Greece/ UC 3.2 Bizkaia, Spain MoSCoW Scale must-have Dependencies Geospatial data availability, Internet Connection Input Needed WGS84 polygon, list of obstacles/no-fly-zones (if applicable) Verification Method Unit and integration testing; input parsing validation Comments Foundational for initiating the mission planning process ID TR-FUN-3.5-2 Name (Optional) Single-UAV Coverage Path Planning Description The service shall compute an optimal coverage path for a single UAV to map/monitor a specified area, considering any obstacles, no-fly zones within the operational area. Category Functional, Design Related Tasks T3.5. Dynamic Resource Allocation for Optimized Safety Surveillance Related Tool UAV-based Path Planning for Coverage Operations Related Use Cases UC 1.2 Nevers, France/ UC 2.1 Karlovac, Croatia/ UC 2.2 Thessaloniki, Greece/ UC 3.2 Bizkaia, Spain MoSCoW Scale must-have Dependencies TR-FUN-3.5-1, UAV availability, specifications and capabilities, Internet Connection @iDriving Consortium – 101147004 Page 34 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Input Needed UAVs characteristics i.e., Battery level, Maximum flight time, Maximum – Minimum altitude for safety and legal constraints, Maximum speed, Payload (camera/FoV) specs Verification Method Simulated mission coverage verification Comments Ensures mission execution for mapping/monitoring an area of interest ID TR-FUN-3.5-3 Name (Optional) Multi-UAV Coverage Path Planning Description The service shall compute cooperative paths for a swarm of UAVs to cooperatively map/monitor a specified area, considering any obstacles, no-fly zones within the operational area. Category Functional, Design Related Tasks T3.5. Dynamic Resource Allocation for Optimized Safety Surveillance Related Tool UAV-based Path Planning for Coverage Operations Related Use Cases UC 1.2 Nevers, France/ UC 2.1 Karlovac, Croatia/ UC 2.2 Thessaloniki, Greece/ UC 3.2 Bizkaia, Spain MoSCoW Scale must-have Dependencies Multiple UAVs, Internet Connection Input Needed Number of UAVs, UAVs characteristics i.e., Battery level, Maximum flight time, Maximum – Minimum altitude for safety and legal constraints, Maximum speed, Payload (camera/FoV) specs Verification Method Simulated mission coverage verification Comments Ensures distributed mission execution across available UAVs for mapping/monitoring an area of interest ID TR-NFUN-3.5-4 Name (Optional) Data Interoperability Description The system shall support standardized data formats for mission inputs/outputs, such as .JSON for area definitions and waypoint extraction. Category Interoperability Related Tasks T3.5. Dynamic Resource Allocation for Optimized Safety Surveillance Related Tool UAV-based Path Planning for Coverage Operations @iDriving Consortium – 101147004 Page 35 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Related Use Cases UC 1.2 Nevers, France/ UC 2.1 Karlovac, Croatia/ UC 2.2 Thessaloniki, Greece/ UC 3.2 Bizkaia, Spain MoSCoW Scale must-have Dependencies Internet Connection Input Needed Data schema definitions Verification Method Integration testing Comments Only the output will be tested through integration testing against the defined data schema. The input will be inserted through the platform that will be developed under Task 3.5, which will act as the interface for mission area and parameter definitions. This requirement operationalizes TR-NFUN-GEN-2 for UAV mission planning. 3.1.2 Requirements related to WP4 ID TR-NFUN-4.1-1 Name (Optional) Jetson NVIDIA Platform Description T4.1 will be implemented on an embedded device, such as the NVIDIA Jetson platform Category Equipment Related Tasks T4.1 - Real-Time Edge Computing for Visual Monitoring Related Tool - Related Use Cases UC1.1/ UC1.2/ U.C. 2.2./U.C 3.2 MoSCoW Scale - Dependencies - Input Needed - Verification Method - Comments - ID TR-FUN-4.1-1 Name (Optional) Seatbelt detection. @iDriving Consortium – 101147004 Page 36 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Description The system shall execute an object detection algorithm on embedded devices connected to roadside cameras or UAVs to identify and classify seat belt usage. Category Functional Related Tasks T4.1 - Real-Time Edge Computing for Visual Monitoring Related Tool Seatbelt detection tool Related Use Cases UC 1.1-Graz, Austria MoSCoW Scale must-have Dependencies Dependence on a dataset containing images of drivers both wearing and not wearing seatbelts for training the deep learning model. Camera's field of view and angle Input Needed Frames or videos with high frame rate Verification Method Training the deep learning model, and performing inference on a provided sample, as part of integration testing Comments Provided data or video samples from different angles are crucial for the functionality of the tool. Most of the images in the current model are captured from road cameras at a 45-degree angle and not from a long distance. Depending on the integration and the placement of the cameras, we may need data to retrain the model from that specific point of view. ID TR-FUN-4.1-2 Name (Optional) Mobile detection. Description The system shall execute an object detection algorithm on embedded devices connected to road cameras or UAVs to identify and classify mobile phone usage. Category Functional Related Tasks T4.1 - Real-Time Edge Computing for Visual Monitoring Related Tool Mobile detection tool Related Use Cases UC 1.1-Graz, Austria/ UC 1.2-Nevers, France MoSCoW Scale must-have Dependencies Dependence on existing datasets for training the deep learning model Camera's field of view and angle @iDriving Consortium – 101147004 Page 37 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Input Needed Frames or videos with high frame rate Verification Method Training the deep learning model, and performing inference on a provided sample, as part of integration testing Comments Provided data or video samples from different angles are crucial for the functionality of the tool. Most of the images in the current model are captured from road cameras at a 45-degree angle and not from a long distance. Depending on the integration and placement of the cameras, we may need data to retrain the model from that specific point of view. In UC 1.2, detecting bikers using mobile phones requires a dedicated dataset. However, we were unable to find any suitable dataset. If such a dataset can be provided, then this tool can also be integrated into UC 1.2. ID TR-FUN-4.1-3 Name (Optional) Helmet Detection Description The system shall execute an object detection algorithm on embedded devices to verify helmet usage by riders of bicycles and motorcycles Category Functional Related Tasks T4.1 - Real-Time Edge Computing for Visual Monitoring Related Tool Helmet detection tool Related Use Cases UC 1.1-Graz, Austria/ UC1.2-Nevers, France MoSCoW Scale must-have Dependencies Dependence on existing datasets for training the deep learning model Camera's field of view and angle Input Needed Frames or videos with high frame rate Verification Method Training the deep learning model, and performing inference on a provided sample, as part of integration testing Comments Provided data or video samples from different angles are crucial for the functionality of the tool. Most of the images in the current model are captured from road cameras at a 45-degree angle and not from a long distance. Depending on the integration and the placement of the cameras, we may need data to retrain the model from that specific point of view. @iDriving Consortium – 101147004 Page 38 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies ID TR-FUN-4.1-4 Name (Optional) License Plate Detection Description The system shall execute an object detection algorithm on embedded devices to identify the license plates of offenders who are not wearing seatbelts or helmets, or who are using mobile phones while driving. Category Functional Related Tasks T4.1 - Real-Time Edge Computing for Visual Monitoring Related Tool License plate detection tool Related Use Cases UC 1.1-Graz, Austria/ UC1.2-Nevers, France MoSCoW Scale should-have Dependencies Dependence on existing datasets for training the deep learning model Camera's field of view and angle Input Needed Frames or videos with high frame rate Verification Method Training the deep learning model, and performing inference on a provided sample, as part of integration testing Comments The model has been trained on a dataset containing close-up images. Additionally, images from street cameras were included to help generalize the model. However, due to the small size of license plates, it may be difficult to detect and read them from drone footage. For accurate license plate recognition, the image must be clear enough to distinguish the characters. ID TR-FUN-4.1-5 Name (Optional) Count Cars Description The system shall detect and count the number of cars present in each video frame captured by connected cameras. Category Functional Related Tasks T4.1 - Real-Time Edge Computing for Visual Monitoring Related Tool Count Car detection tool Related Use Cases UC 1.1-Graz, Austria, UC2.2-Thessaloniki, Greece MoSCoW Scale should-have Dependencies Dependence on existing datasets for training the deep learning model @iDriving Consortium – 101147004 Page 39 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Camera's field of view and angle Provision of data Input Needed Frames or videos with high frame rate Verification Method Training the deep learning model, and performing inference on provided samples, as part of integration testing Comments This tool was added later, following the plenary meeting in Graz. Provided data or video samples from different angles are crucial for the functionality of the tool. Most of the images in the current model are captured from road cameras at a 45-degree angle and not from a long distance. Depending on the integration and the placement of the cameras, we may need data to retrain the model from that specific point of view. ID TR-FUN-4.1-6 Name (Optional) Zebra crossing detection Description The system shall detect the presence of zebra crossings in video frames captured by road cameras or UAVs. Category Functional Related Tasks T4.1 - Real-Time Edge Computing for Visual Monitoring Related Tool Zebra Crossing detection tool Related Use Cases UC 1.2. - Nevers, France MoSCoW Scale should-have Dependencies Dependence on existing datasets for training the deep learning model Camera's field of view and angle Provision of data Input Needed Frames or videos with high frame rate Verification Method Training the deep learning model, and performing inference on provided samples, as part of integration testing Comments This tool was added following a bilateral meeting with TEKNIKER for T4.2. The purpose is to determine whether a person is walking on a zebra crossing or not – similar to checking if someone is walking on sidewalk. If needed CERTH will provide only the coordinates of the zebra crossing, by combining the zebra crossing coordinates within the frame with the detected position of the person. We can assess @iDriving Consortium – 101147004 Page 40 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies whether the individual is walking on the street or within the designed crossing area. Sidewalk detection is not feasible, as most pavements require polygon-shaped annotations to be accurately bounded. However, our current models only support rectangular bounding boxes. Training a model to handle both rectangular and polygon annotation is not supported within our framework, Therefore, only zebra crossing can be detected, given that the provided dataset includes rectangular bounding box annotations. ID TR-FUN-4.1-7 Name (Optional) Fallen Tree and Rockslide detection tool Description The system shall execute an object detection algorithm on embedded devices to monitor landscapes and roadways for fallen trees and rockslides using video feeds from UAVs. It analyses aerial imagery to quickly identify and assess potential hazards. Category Functional Related Tasks T4.1 - Real-Time Edge Computing for Visual Monitoring Related Tool Disaster detection tool Related Use Cases UC2.2 - Thessaloniki, Greece MoSCoW Scale must-have Dependencies Dependence on existing datasets for training the deep learning model Camera's field of view and angle Provision of data Input Needed Frames or videos with high frame rate Verification Method Training the deep learning model, and performing inference on provided samples, as part of integration testing Comments It is important to provide us with data from the field of interest that includes the relevant objects of interest ID TR-FUN-4.1-8 Name (Optional) Crashed car detection tool Description The system shall execute an object detection algorithm on embedded devices to identify crashed vehicles in real time using video feeds from CCTV cameras or UAVs. It analyses the imagery to detect unusually stationary vehicles or collisions on the road and promptly alerts emergency services. @iDriving Consortium – 101147004 Page 41 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Category Functional Related Tasks T4.1 - Real-Time Edge Computing for Visual Monitoring Related Tool Crashed Car detection tool Related Use Cases UC 3.2 - Bizkaia, Spain MoSCoW Scale must-have Dependencies Dependence on existing datasets for training the deep learning model Camera's field of view and angle Provision of data Input Needed Frames or videos with high frame rate Verification Method Training the deep learning model, and performing inference on provided samples, as part of integration testing Comments It is important to provide us with data from the relevant field that includes the objects of interest. ID TR-FUN-4.2-1 Name (Optional) Red light violation detector Description Detect vehicles in real-time and determine if they cross a virtual stop line during a red traffic light phase. Category Functional Related Tasks Task 4.2 Related Tool Red Light Violation Detector Related Use Cases UR_UC1.1_17 MoSCoW Scale Should-have Dependencies Real-time video input stream, accurate traffic light state data, correct virtual line configuration. Input Needed Video stream from surveillance cameras, real-time traffic light status data. Obtaining this information does not appear to be feasible. It cannot be acquired through computer vision, and it is currently unclear whether access to the traffic light timing data will be possible for the intended use case. Verification Method System validation with test video sequences including known violation cases; cross-verification of alerts and captured evidence (image + timestamp + traffic light status). @iDriving Consortium – 101147004 Page 48 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Related Tasks TASK 4.4: Smart Environmental Condition Monitoring for Proactive Road Safety Measures Related Tool Tool 4.4.2: Weather Prediction Tool using WRF and Data Assimilation Related Use Cases UC2.2 - Thessaloniki, Greece / UC 2.1 Karlovac, Croatia MoSCoW Scale Must-have Dependencies Git, docker, Web Framework (like Django Rest Framework), Relational Database, either locally installed Open Source LLM or Cloud Provided LLM (Accessed via API), Python Input Needed Station Real Time Data, Forecast Data, Use Case data Verification Method User Feedback Comments - ID TR-FUN-4.5-1 Name (Optional) 3D Reconstruction of Road Environments Description The tool transforms 2D visual data captured by UAVs or groundbased moving cameras in road environments into 3D representations, enhancing safety, enabling remote inspections, and providing valuable maintenance insights. The scale of the scene extended from a single defect in a small-scale scene to a larger partition of the road representing broader environments. The reconstruction method must be adapted according to the scale of the scene, whether it is a localized defect or a larger road segment. Also, the tool reconstructs scenes with enriched information, enhancing scene understanding and highlighting potential road defects in 3D space. Category Functional, Design Related Tasks T4.5 AI-driven 3D Scene Generation for Safety & Maintenance Monitoring Related Tool 3D-Smart Tool Related Use Cases UC 2.1 Karlovac, Croatia MoSCoW Scale should-have Dependencies • High quality visual data acquired from UAVs or groundbased moving cameras. • Input from Task T4.3 will be used for defect identification to apply multiresolution capabilities in the region surrounding detected anomalies @iDriving Consortium – 101147004 Page 49 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Input Needed • High-resolution video or sequential high-resolution images. • Input from Task T4.3 Verification Method Quality metrics against users' perception, visual fidelity against ground truth, rendering quality metrics, projection errors Comments The tool is highly dependent on the data capture. The entire target area must be thoroughly covered from multiple viewpoints to ensure accurate reconstruction. In the case of ground-based moving capture, the camera should remain as stable as possible while in motion. Extracted images should have significant overlap (ideally 70%). The output will be visualized according to the method that will be used. Possible options: Local browser link for interactive viewing Export of 3D Gaussians to .ply file for external tools Method’s integrated SIBR viewer ID TR-FUN-4.5-2 Name (Optional) 3D Reconstruction of Car Accident Description The tool transforms 2D visual data captured by UAVs at potential car accidents scenes into 3D replications, enhancing safety, enabling remote inspections, and providing valuable incident monitoring. Category Functional, Design Related Tasks T4.5 AI-driven 3D Scene Generation for Safety & Maintenance Monitoring Related Tool 3D-Smart Tool Related Use Cases UC 3.2 Bizkaia, Spain MoSCoW Scale should-have Dependencies High quality visual data acquired from UAVs or ground-based moving cameras. Input Needed High-resolution video or sequential high-resolution images. Verification Method Quality metrics against users' perception, visual fidelity against ground truth, rendering quality metrics, projection errors Comments The tool is highly dependent on the data capture. The entire target area must be thoroughly covered from multiple viewpoints to ensure accurate reconstruction. In the case of ground-based @iDriving Consortium – 101147004 Page 50 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies moving capture, the camera should remain as stable as possible while in motion. Extracted images should have significant overlap (ideally 70%). The output will be visualized according to the method that will be used. Possible options: Local browser link for interactive viewing Export of 3D Gaussians to .ply file for external tools Method’s integrated SIBR viewer 3.1.3 Requirements related to WP5 ID TR-FUN-5.1-1 Name (Optional) Data integration tool Description The systems must take account multi-source data from IDriving monitored network. Data must be adapted and process for continuous learning of KPI defined in SCC Category Design, interoperability Related Tasks Dynamic Update of Criteria Catalogue Through Continuous Learning Related Tool Dynamic monitoring tool Related Use Cases UC1.1, UC1.2, UC2.1, UC2.2, UC3.1, UC3.2 MoSCoW Scale must-have Dependencies Safety Criteria Catalogue Input Needed Data from UCs relative to KPI define in SCC (data frequency, quality, or minima) Verification Method Integration testing, functional testing Comments - ID TR-FUN-5.1-2 Name (Optional) KPI computation and storage Description The system must compute KPI and learn indicators characteristics based on infrastructure and historical measurements. KPI are defined inside a SCC python database. The results are stored for evaluation, learning and visualisation. Category Functional @iDriving Consortium – 101147004 Page 51 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Related Tasks Dynamic Update of Criteria Catalogue Through Continuous Learning Related Tool Dynamic monitoring tool Related Use Cases UC1.1, UC1.2, UC2.1, UC2.2, UC3.1, UC3.2 MoSCoW Scale must-have Dependencies Safety Criteria Catalogue Input Needed Historical data from monitored areas Verification Method Integration testing, functional testing Comments - ID TR-FUN-5.1-3 Name (Optional) KPI monitoring dashboard Description The system shall provide a dashboard for visualizing KPI time series, supporting multiple visualization formats such as maps, heatmaps, and graphs. Category Functional Related Tasks Dynamic Update of Criteria Catalogue Through Continuous Learning Related Tool Dynamic monitoring tool Related Use Cases UC1.1, UC1.2, UC2.1, UC2.2, UC3.1, UC3.2 MoSCoW Scale must-have Dependencies Internet connection (API) Input Needed KPI data Verification Method Integration testing, functional testing Comments - ID TR-FUN-5.1-4 Name (Optional) KPI Continuous learning Description The system shall support forecasting and evaluation of KPIs based on incoming data streams. Category Functional Related Tasks Dynamic Update of Criteria Catalogue Through Continuous Learning Related Tool Dynamic monitoring tool @iDriving Consortium – 101147004 Page 52 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Related Use Cases UC1.1, UC1.2, UC2.1, UC2.2, UC3.1, UC3.2 MoSCoW Scale must-have Dependencies Input Needed KPI data, Traffic data Verification Method Integration testing, functional testing Comments - ID TR-FUN-5.3-1 Name (Optional) Data integration to digital twins Description The systems must take account multi-source data from IDriving monitored network. Data must be adapted and process to send to digital twin and store to perform offline calibration of digital twin simulation tools Category Functional, interoperability Related Tasks Digital Twin Powered Predictive Safety Measures and Warning Systems Related Tool SUMO/CARLA Related Use Cases UC1.1, UC1.2, UC2.2, UC3.1, UC3.2 MoSCoW Scale must-have Dependencies - Input Needed Traffic Data from UCs Data/events from ground camera or UAVs Weather data for UC2.2 Verification Method Integration testing, functional testing Comments - ID TR-FUN-5.3-2 Name (Optional) Digital twin calibration Description The system shall build a digital twin of road networks in monitored areas to simulate road user behaviour based on real data. The resulting models shall be stored and made available for reuse within the iDriving project. Category Functional Related Tasks Digital Twin Powered Predictive Safety Measures and Warning Systems @iDriving Consortium – 101147004 Page 53 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Related Tool SUMO/CARLA, T3.2 Related Use Cases UC1.1, UC1.2, UC2.2, UC3.1, UC3.2 MoSCoW Scale must-have Dependencies - Input Needed Traffic Data from UCs Verification Method Integration testing, functional testing Comments - ID TR-FUN-5.3-3 Name (Optional) Multi road user simulation Description The system shall use a simulation-based digital twin to perform traffic prediction based on real-time and historical data. It shall generate user trajectories under various conditions and configurations. Category Functional Related Tasks Digital Twin Powered Predictive Safety Measures and Warning Systems Related Tool SUMO Related Use Cases UC1.1, UC1.2, UC2.2, UC3.1, UC3.2 MoSCoW Scale must-have Dependencies - Input Needed Traffic data, SUMO calibrated models Verification Method Integration testing, functional testing Comments - ID TR-FUN-5.3-4 Name (Optional) Intersection safety simulation Description The system shall perform a 3D simulation of critical area of the network (e.g. intersections) In order to evaluate risk for roadusers. Category Functional Related Tasks Digital Twin Powered Predictive Safety Measures and Warning Systems Related Tool CARLA @iDriving Consortium – 101147004 Page 54 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Related Use Cases UC1.1, UC1.2, UC2.2, UC3.1, UC3.2 MoSCoW Scale must-have Dependencies - Input Needed Traffic data, SUMO calibrated models Verification Method Integration testing, functional testing Comments - ID TR-FUN-5.3-5 Name (Optional) Safety analysis and communication system Description From simulation data, provide alerts and safety measures for targeted network for iDriving prototype. Alerts are stored to perform safety assessment and tool evaluation. Category Functional Related Tasks Digital Twin Powered Predictive Safety Measures and Warning Systems Related Tool SUMO/CARLA Related Use Cases UC1.1, UC1.2, UC2.2, UC3.1, UC3.2 MoSCoW Scale must-have Dependencies - Input Needed Traffic data, SUMO calibrated models Verification Method Integration testing, functional testing Comments - ID TR-FUN-5.4-1 Name (Optional) Infrastructure Segment Risk Calculation Description The system shall calculate the dynamic risk for each infrastructure segment using both historical and real-time data, applying an FMEA-based risk assessment methodology. The risk calculation must consider road characteristics, traffic flow, and structural vulnerability information. This risk evaluation enables simulation of what-if scenarios and supports both the definition of maintenance needs and the planning of optimal maintenance interventions. Category Functional Related Tasks T5.4 @iDriving Consortium – 101147004 Page 55 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Related Tool Dynamic Risk Assessment Tool Related Use Cases UC2.1_4 (Karlovac), UC2.1_5 MoSCoW Scale Must-have Dependencies Requires input from condition monitoring tools (detected structural vulnerabilities), traffic models and road characteristic data. Data integration with Tool 1.2 (Logistics Planning Tool) Input Needed Real-time road characteristics, traffic flow data, vulnerability data and scenario simulation results (optionally including damage information) Verification Method Validation of calculated risk metrics against expert assessments or field data; Functional tests verifying real-time risk updates when input conditions (e.g., traffic or vulnerability levels) change; Review of maintenance recommendations and plans generated based on risk indicators Comments This requirement is a key enabler for proactive maintenance strategies based on Digital Twin architecture. It ensures risk calculations are dynamically adapted to evolving road and traffic conditions and provides essential input to logistics and maintenance planning tools. ID TR-FUN-5.4-2 Name (Optional) Dynamic calculation of Health Index / Road Condition Index (RCI) Description The system must dynamically compute Health Index or Road Condition Index values based on real-time (from condition monitoring and detecting road defects) and forecasted data, weather, and traffic. Category Functional Related Tasks T5.4 Related Tool Health and Logistic Management Tool Related Use Cases UC2.1_4 (Karlovac) MoSCoW Scale Must Have Dependencies Integration with sensor-based data acquisition systems— primarily imaging devices—for road surface inspection; execution of computer vision algorithms to detect and classify road surface defects; and incorporation of weather and traffic forecasting pipelines to enable dynamic health index computation @iDriving Consortium – 101147004 Page 56 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Input Needed Real-time or near real-time monitoring and processing of road condition data to estimate the infrastructure's health status, complemented by meteorological and traffic data to enhance the statistical accuracy of the health index estimation. Verification Method Comparison against actual condition measurements in pilot scenarios or simulated outcomes form Digital Twins model Comments Critical for enabling proactive and cost-effective maintenance. Enables downstream planning and risk analysis via connected tools (Module 1.1 and Module 1.3) ID TR-FUN-5.4-3 Name (Optional) Maintenance Scheduling and Optimization Description The system must generate shortand long-term maintenance plans based on the dynamic risk, health indicators, traffic flow forecasts, weather conditions, and available resources. It must balance risk reduction, maintenance costs, and resource optimization by applying metaheuristic optimization algorithms and predefined operational constraints. Category Functional Related Tasks T5.4 Related Tool Maintenance Scheduling Tool Related Use Cases UC2.1_4 (Karlovac), UC2.1_5 MoSCoW Scale Must-have Dependencies Requires input from the Dynamic Risk Assessment Tool (Module 1.1) and Health and Logistics Management Tool (Module 1.2); Requires access to operational maintenance plans and constraints databases; Depends on real-time updates for traffic and weather forecasts. Input Needed Risk metrics, health indicators (RCI/Health Index), traffic forecasts, weather forecasts, predefined maintenance plans, available resource information Verification Method Validation through simulation of generated maintenance plans against expected optimization goals (cost vs risk trade-offs); Review of short-term and long-term plans to verify correct prioritization and resource assignment; Testing of exporting functionality (CSV/JSON formats). Comments Critical to ensure that maintenance interventions are not only riskdriven but also optimized for costs and operational feasibility. The tool must allow dynamic re-scheduling in case of updated risks or unexpected traffic/weather events. It is a key output component for @iDriving Consortium – 101147004 Page 57 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies supporting road managers in real-time and strategic decisionmaking. ID TR-FUN-5.5-1 Name (Optional) Interface for Digital-Twin based control centre Description The system shall provide a user interface for the Digital Twin-based Control Center, compatible with both desktop applications and XR devices. Category Functional, Design, interoperability, security, regulatory/compliance Related Tasks T5.5 Digital Twin-Based Control Center with XR features for Enhanced Situational Awareness Related Tool Digital-Twin based control centre Related Use Cases UC1.1, UC1.2, UC2.1, UC2.2, UC3.1, UC3.2 MoSCoW Scale Must-have Dependencies Interface depends on the device performance Data received from the other components (centralized at server level) Input Needed - User requirements list Verification Method Usability testing Comments - ID TR-FUN-5.5-2 Name (Optional) Alerting and communication service Description The system shall include a dedicated service for bidirectional communication with the main server, handling the sending and receiving of messages, including alerts and notifications. Category Functional, Performance, interoperability, security, regulatory/compliance Related Tasks T5.5 Digital Twin-Based Control Centre with XR features for Enhanced Situational Awareness Related Tool Digital-Twin based control centre Related Use Cases UC1.1, UC1.2, UC2.1, UC2.2, UC3.1, UC3.2 MoSCoW Scale Must-have Dependencies Internet Connection, External services (e.g. weather information) @iDriving Consortium – 101147004 Page 64 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies UR_UC1.2_31 TR-FUN-3.3-2, TR-FUN-4.1-6, TR-FUN-4.2-3, TR-FUN-5.5-2, TR-FUN-5.3-5 UR_UC1.2_32 TR-FUN-3.3-2, TR-FUN-4.2-4, TR-FUN-5.5-2, TR-FUN-5.3-5 UR_UC2.1_1 TR-FUN-4.3-1, TR-FUN-4.3-2, TR-FUN-5.4-1 UR_UC2.1_2 TR-FUN-5.1-3, TR-FUN-4.3-1, TR-FUN-4.3-2 UR_UC2.1_3 TR-FUN-5.5-3 UR_UC2.1_4 TR-FUN-5.4-1 UR_UC2.1_5 TR-FUN-5.4-3 UR_UC2.1_6 TR-FUN-5.5-3 UR_UC2.1_7 TR-FUN-5.5-1 UR_UC2.1_8 TR-FUN-3.2-2, TR-FUN-5.4-1 UR_UC2.1_9 TR-FUN-5.5-2 UR_UC2.1_10 TR-FUN-5.5-2 UR_UC2.1_11 TR-FUN-5.5-2 UR_UC2.1_12 TR-FUN-3.5-1, TR-FUN-3.5-2, TR-FUN-3.5-3, TR-NFUN-3.5-4, TR-FUN-4.2-2 UR_UC2.1_13 TR-FUN-3.5-1, TR-FUN-3.5-2, TR-FUN-3.5-3, TR-NFUN-3.5-4, TR-FUN-4.2-3 UR_UC2.1_14 TR-FUN-3.5-1, TR-FUN-3.5-2, TR-FUN-3.5-3, TR-NFUN-3.5-4 UR_UC2.1_15 TR-FUN-5.5-1 UR_UC2.1_17 TR-FUN-5.5-2 UR_UC2.1_19 TR-FUN-4.5 UR_UC2.2_1 TR-FUN-3.3-2 UR_UC2.2_2 TR-FUN-3.3-2 UR_UC2.2_3 TR-FUN-3.3-2 UR_UC2.2_4 TR-FUN-3.3-2 UR_UC2.2_5 TR-FUN-3.3-2 UR_UC2.2_6 TR-FUN-3.3-1 UR_UC2.2_7 TR-FUN-3.3-1 UR_UC2.2_8 TR-FUN-3.3-1 UR_UC2.2_9 TR-FUN-3.3-1 @iDriving Consortium – 101147004 Page 65 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies UR_UC2.2_10 TR-FUN-3.2-1 UR_UC2.2_11 TR-FUN-3.2-1 UR_UC2.2_12 TR-FUN-3.3-2 UR_UC2.2_13 TR-FUN-3.3-1 UR_UC2.2_14 TR-FUN-3.3-2 UR_UC2.2_15 TR-FUN-3.2-1 UR_UC2.2_16 TR-FUN-3.2-1 UR_UC2.2_17 TR-FUN-5.5-3 UR_UC2.2_18 TR-FUN-5.5-1 UR_UC2.2_19 TR-FUN-5.5-1 UR_UC2.2_20 TR-FUN-5.5-3 UR_UC2.2_23 TR-FUN-5.5-3 UR_UC2.2_24 TR-FUN-5.5-3 UR_UC2.2_25 TR-FUN-5.5-3 UR_UC2.2_26 TR-FUN-5.5-3 UR_UC2.2_27 TR-FUN-4.4-1, TR-FUN-4.4-5 UR_UC2.2_28 TR-FUN-4.4-1, TR-FUN-4.4-5 UR_UC2.2_29 TR-FUN-4.4-1, TR-FUN-4.4-5 UR_UC2.2_30 TR-FUN-4.4-1, TR-FUN-4.4-5 UR_UC2.2_31 TR-FUN-4.4-1, TR-FUN-4.4-5 UR_UC2.2_32 TR-FUN-4.4-1, TR-FUN-4.4-5 UR_UC2.2_33 TR-FUN-3.5-1, TR-FUN-3.5-2, TR-FUN-3.5-3, TR-NFUN-3.5-4, TR-FUN-4.1-7 UR_UC2.2_34 TR-FUN-3.5-1, TR-FUN-3.5-2, TR-FUN-3.5-3, TR-NFUN-3.5-4, TR-FUN-4.1-7 UR_UC2.2_35 TR-FUN-3.5-1, TR-FUN-3.5-2, TR-FUN-3.5-3, TR-NFUN-3.5-4, TR-FUN-4.1-7 UR_UC2.2_36 TR-FUN-3.5-1, TR-FUN-3.5-2, TR-FUN-3.5-3, TR-NFUN-3.5-4 UR_UC3.1_1 TR-FUN-4.2-4 UR_UC3.1_2 TR-FUN-4.2-4 UR_UC3.1_3 TR-FUN-4.2-2, TR-FUN-4.2-4 @iDriving Consortium – 101147004 Page 66 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies UR_UC3.1_6 TR-FUN-3.3-2 UR_UC3.1_7 TR-FUN-3.3-1, TR-FUN-3.3-2, TR-NFUN-3.4.4, TR-FUN-5.5-2 UR_UC3.1_8 TR-FUN-3.3-1 UR_UC3.1_13 TR-FUN-5.3-5 UR_UC3.1_14 TR-FUN-3.3-1 UR_UC3.1_16 TR-FUN-3.3-2, TR-FUN-5.3-5, TR-FUN-5.5-2 UR_UC3.1_18 TR-FUN-3.3-1, TR-FUN-3.3-2 UR_UC3.1_21 TR-FUN-5.1-3, TR-FUN-5.5-1, TR-FUN-5.5-3 UR_UC3.1_22 TR-FUN-5.1-2, TR-FUN-5.1-3, TR-FUN-5.5-3 UR_UC3.1_23 TR-FUN-3.3-2, TR-NFUN-3.4.4, TR-FUN-5.5-2 UR_UC3.1_26 TR-NFUN-GEN-9 UR_UC3.1_27 TR-FUN-3.3-1 UR_UC3.1_28 TR-FUN-5.3-5 UR_UC3.2_1 TR-FUN-3.2-1, TR-FUN-3.3-1, TR-FUN-3.3-2 UR_UC3.2_2 TR-FUN-3.3-2 UR_UC3.2_3 TR-FUN-3.3-1, TR-FUN-3.3-2 UR_UC3.2_4 TR-FUN-3.3-1 UR_UC3.2_5 TR-FUN-3.2-1, TR-FUN-3.3-1 UR_UC3.2_6 TR-FUN-3.2-1, TR-FUN-3.3-2 UR_UC3.2_7 TR-FUN-4.1-8 UR_UC3.2_8 TR-FUN-4.1-8, TR-FUN-4.2-4 UR_UC3.2_9 TR-FUN-3.5-1, TR-FUN-3.5-2, TR-FUN-3.5-3, TR-NFUN-3.5-4 UR_UC3.2_10 TR-FUN-4.5 UR_UC3.2_11 TR-FUN-5.1-2, TR-FUN-5.1-3, TR-FUN-5.1-4 UR_UC3.2_12 TR-FUN-3.3-1, TR-FUN-3.3-2, TR-FUN-4.1-8 UR_UC3.2_13 TR-FUN-3.3-1 UR_UC3.2_14 TR-FUN-3.3-2 @iDriving Consortium – 101147004 Page 67 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies 5 Overall Architecture & Components 5.1 Methodology for Defining the Architecture The iDriving architecture adopts the C4 Model, a widely recognized methodology for the visualization and communication of complex software systems. This approach enables clear, layered representations that address different stakeholder concerns, ranging from the overall business context to detailed technical designs. The C4 Model, specifically selected for this deliverable, provides the following architectural perspectives: • Context View: Illustrates the system as a “black box” in its environment, identifying all major external actors and their interactions with iDriving. • Container View: Breaks down iDriving into its main technology containers (applications, databases, services), focusing on the responsibilities and integration points between them for each use case. • Component View: (Detailed per-container) Provides a further breakdown of how key containers are internally structured and how their sub-components collaborate to realize functional requirements. • Code View: (Omitted from this deliverable) Reserved for implementationlevel details. This layered approach ensures all audiences—technical, managerial, and operational—can clearly understand the system at the right level of abstraction for their needs. Furthermore, the model’s modularity allows us to articulate both common patterns across use cases and bespoke design elements that address unique local challenges. 5.2 iDriving Context View The Context View (see Figure 2) presents a high-level overview of the iDriving ecosystem, visualizing its position within the wider operational and stakeholder landscape. This diagram identifies the main categories of external actors—Traffic Authorities / Road Operators, UAV Operators / Maintenance Teams, and Road Users—and delineates their primary data exchanges with the iDriving system. Purpose: The Context View establishes the project’s system boundaries, clarifies roles and responsibilities, and sets the stage for understanding subsequent architectural details. It ensures that all project stakeholders—from technical teams to management and end users—share a mutual understanding of who interacts with iDriving and what the essential information flows are. @iDriving Consortium – 101147004 Page 68 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Figure 2 - iDriving Context View 5.3 iDriving Containers View The Container View diagrams provide a focused breakdown of the iDriving system into its major functional and technological building blocks (“containers”). These diagrams capture the core architecture of iDriving, presenting a layered representation that organizes system components according to their roles in the overall workflow: • Data Acquisition: Edge devices, sensors, UAVs, and other sources responsible for gathering raw data from the field. • Detection & Prediction: AI/ML tools and visual analytics modules that transform raw data into actionable insights and early warnings. • Core Back-End & Optimization: Centralized services, orchestration engines, and simulation tools, which process, store, and optimize system outputs. • User-Facing & Control: Applications, dashboards, and control centres that deliver insights and instructions to end users and operators. Colour Coding and Layering: Each container in the diagram is color-coded to reflect its alignment with specific Tasks, ensuring immediate visual clarity regarding component ownership and functional grouping. Layered structuring (vertical separation) emphasizes the flow of data and control from acquisition through to actionable outcomes. @iDriving Consortium – 101147004 Page 69 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Purpose: The Containers View enables all stakeholders to grasp briefly and immediately, which software and hardware modules are active in a given use case, how data moves between them, and how the overall system achieves its functional and nonfunctional requirements. This approach facilitates traceability from user needs to technical solutions, supports impact analysis for changes, and provides a strong basis for further detailed design in the Components’ View. 5.3.1 Containers View – Consolidated This section provides a unified Container View diagram that aggregates the common architectural elements across all use cases. It highlights the shared infrastructure, core backend services, and cross-cutting components (e.g., Dataspace Connector, Kafka broker) that underpin the iDriving ecosystem. This consolidated view supports system-wide integration and reuse. @iDriving Consortium – 101147004 Page 70 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Figure 3 - Consolidated iDriving Containers view Figure 4 - Containers View Colour Legend @iDriving Consortium – 101147004 Page 71 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies 5.3.2 Containers View Per Use Case While the consolidated view outlines an overall logical view of the system, each pilot deployment has its own architectural nuances. The following sub-sections present use-case-specific Container View diagrams, showcasing localized deployment patterns, data flows, and components tailored to the particular context and objectives of each site. 5.3.2.1 Use case 1.1 Graz, Austria Container View diagram for UC 1.1 Graz is presented in Figure 5. Figure 5 - Use Case 1.1 Containers View @iDriving Consortium – 101147004 Page 72 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies 5.3.2.2 Use Case 1.2 Nevers, France Container View diagram for UC 1.2 Nevers is presented in Figure 6. Figure 6 - Use Case 1.2 Containers View @iDriving Consortium – 101147004 Page 73 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies 5.3.2.3 Use Case 2.1 Karlovac, Croatia Container View diagram for UC 2.1 Karlovac is presented in Figure 7. Figure 7 - Use Case 2.1 Containers View @iDriving Consortium – 101147004 Page 80 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Architectural Diagram and Subcomponents (components View) @iDriving Consortium – 101147004 Page 81 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Expected TRL TRL Technologies The following technologies will be used for the implementation of the component: • SUMO traffic simulation software Deployment Premises The testing of the various scenarios defined will be done in simulated environment and/or in real-time implementation. The route guidance tool will be an add-on module for SUMO traffic simulator. The system will be tested virtually on servers within the partners’ premises. Involved Partners MBL; TUC Related Technical requirements TR-FUN-3.2-1 @iDriving Consortium – 101147004 Page 82 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Table 11 - Signal Control Tool – Architectural Design and Description Signal Control Tool: Safety-Optimized Traffic Management Using Intelligent Algorithms for Diverse Vehicle Flows Description The system receives information from different sources (cameras, induction loop detectors, magnetometer sensors etc.). The signal control tool provides dynamic adjustment of traffic light timings for an identified intersection(s) based on the actual traffic information coming from the neighbouring links of the intersection. At each signal cycle, measurements or estimates of queue lengths or accumulated vehicles are feeding the control strategy, whose main goal is to improve efficiency of the trips served by the controlled intersection(s). The optimal signal timings are computed in real-time and broadcasted to the traffic lights. @iDriving Consortium – 101147004 Page 83 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Architectural Diagram and Subcomponents (components View) Expected TRL TRL Technologies The following technologies will be used for the implementation of the Signal Control Component: • SUMO traffic simulation software @iDriving Consortium – 101147004 Page 84 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Deployment Premises The testing of the various scenarios defined will be done in a simulated environment and/or in real-time implementation. The signal control tool will be an add-on module for the SUMO traffic simulator. The system will be tested virtually on servers within the partners’ premises. Involved Partners MBL; TUC Related Technical requirements TR-FUN-3.2-2 @iDriving Consortium – 101147004 Page 85 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Table 12 - Mobile Application – Architectural Design and Description Mobile Application: Mobile and in-vehicle applications for early warnings Description A user-centric software system that serves both drivers and non-driving road users. It connects directly to user’s mobile device (Android system) and the Digital Twin to fetch real-time data on road conditions, traffic, and other potential hazards Architectural Diagram and Subcomponents (components View) @iDriving Consortium – 101147004 Page 86 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Expected TRL TRL6 Technologies The following technologies will be used for the implementation of the X Component: • Unity • C# • REST API • Photoshop • Blender 3D Deployment Premises Android based mobile device with WiFi connection Involved Partners TUC Related Technical requirements TR-FUN-3.3-1, TR-FUN-3.3-2 Table 13 - In-Vehicle Application – Architectural Design and Description In-Vehicle Application: Mobile and in-vehicle applications for early warnings Description A user-centric software system that serves drivers. It connects directly to the vehicle’s built in infotainment system (Android system) and the Digital Twin to fetch real-time data on road conditions, traffic, and other potential hazards @iDriving Consortium – 101147004 Page 87 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Architectural Diagram and Subcomponents (components View) Expected TRL TRL6 Technologies The following technologies will be used for the implementation of the X Component: • Unity • C# • REST API @iDriving Consortium – 101147004 Page 88 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies • Photoshop • Blender 3D Deployment Premises In-vehicle Android system with WiFi connection Involved Partners TUC Related Technical requirements TR-FUN-3.3-1, TR-FUN-3.3-2 @iDriving Consortium – 101147004 Page 89 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Table 14 - Aerial Surveillance UAV System – Architectural Design and Description Aerial Surveillance UAV system: Aerial Surveillance in Incident Management and Maintenance Tasks Description A UAV system will be delivered, equipped with the ability to host different sensors and cameras (i.e. high-resolution camera, thermal etc.) in order to ensure precise, efficient, effective data collection from the areas of interest on demand and/or near real time and address the iDriving specific needs in trials. The UAV will be able to perform manual or autonomous flights by preparing a flight plan prior to the mission. For the pilot use cases and in the case of the need to execute the AI Algorithms in near real-time, the following two approaches are available: 1. A NVIDIA AI-oriented computer will be available for mounting on the UAV where the payloads will interact with the onboard computer to achieve the objectives of each specific mission according to PUC requirements. 2. Transfer the video stream in real time in the ground station computer where the AI algorithms will be installed, executed, and then transfer the required data to any cloud/server needed. Architectural Diagram and Subcomponents (components View) @iDriving Consortium – 101147004 Page 96 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Technologies The following technologies will be used for the implementation of the X Component: • Python • Deep learning detection models Deployment Premises Embedded on edge device Involved Partners MBL, UNI.EIFFEL Related Technical requirements TR-FUN-4.1-5 Table 18 - License Plate Detection Tool – Architectural Design and Description License Plate Detection Tool: Real-Time Edge Computing for Visual Monitoring Description This tool runs on embedded devices and uses computer vision algorithms to automatically identify license plates from video footage once a traffic violation is detected. After recognizing offenses such as seat belt non-compliance or cell phone usage while driving, the tool accurately captures the license plate of the offending vehicle. @iDriving Consortium – 101147004 Page 97 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Architectural Diagram and Subcomponents (components View) Expected TRL TRL 6 Technologies The following technologies will be used for the implementation of the X Component: • Python • Deep learning detection models Deployment Premises The tool will run on an embedded device such as the NVIDIA Jetson Involved Partners SIMAVI, TEKNIKER, UNI.EIFFEL Related Technical requirements TR-FUN-4.1-4, @iDriving Consortium – 101147004 Page 98 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Table 19 - Helmet Detection Tool – Architectural Design and Description Helmet Detection Tool: Real-Time Edge Computing for Visual Monitoring Description This tool leverages advanced computer vision algorithms to automatically detect helmet usage by motorcyclists or cyclists in realtime. It can run on embedded devices connected to road cameras or UAVs (drones), enabling effective monitoring of helmet compliance in various environments. By ensuring riders wear helmets, the tool promotes safety and supports traffic regulation enforcement. Architectural Diagram and Subcomponents (components View) Expected TRL TRL 6 Technologies The following technologies will be used for the implementation of the X Component: • Python • Deep learning detection models @iDriving Consortium – 101147004 Page 99 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Deployment Premises The tool will run on an embedded device such as the NVIDIA Jetson Involved Partners SIMAVI, TEKNIKER, UNI.EIFFEL Related Technical requirements TR-FUN-4.1-3 Table 20 - Zebra Crossing Detection Tool – Architectural Design and Description Zebra Crossing Detection Tool: Real-Time Edge Computing for Visual Monitoring Description The system shall detect the presence of zebra crossing in video frames captured by roadside cameras or UAVs Architectural Diagram and Subcomponents (components View) Expected TRL TRL 6 @iDriving Consortium – 101147004 Page 100 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Technologies The following technologies will be used for the implementation of the X Component: • Python • Deep learning detection models Deployment Premises The tool will run on an embedded device such as the NVIDIA Jetson Involved Partners TEKNIKER Related Technical requirements TR-FUN-4.1-6 Table 21 - Fallen Tree and Rockslide Detection Tool – Architectural Design and Description Fallen tree and Rockslide Plate Detection Tool: Real-Time Edge Computing for Visual Monitoring Description This tool utilizes computer vision algorithms to monitor landscapes and roadways for fallen trees and rockslides using video feeds from UAVs (drones). It analyses aerial imagery to quickly identify and assess potential hazards, enabling prompt reporting and response. @iDriving Consortium – 101147004 Page 101 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Architectural Diagram and Subcomponents (components View) Expected TRL TRL 6 Technologies The following technologies will be used for the implementation of the X Component: • Python • Deep learning detection models Deployment Premises The tool will run on an embedded device such as the NVIDIA Jetson Involved Partners MBL, UNI.EIFFEL, ACCELI Related Technical requirements TR-FUN4.1-7 @iDriving Consortium – 101147004 Page 102 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Table 22 - Crashed Vehicle Detection Tool – Architectural Design and Description Crashed vehicle Detection Tool: Real-Time Edge Computing for Visual Monitoring Description This tool employs computer vision algorithms to identify crashed vehicles in real-time using video feeds from CCTV cameras or UAVs (drones). It analyses the imagery to detect unusual stationary vehicles or collisions on the road, promptly alerting emergency services. Architectural Diagram and Subcomponents (components View) Expected TRL TRL 6 Technologies The following technologies will be used for the implementation of the X Component: • Python • Deep learning detection models Deployment Premises The tool will run on an embedded device such as the NVIDIA Jetson Involved Partners ACCELI, UNI.EIFFEL, MBL, TEKNIKER @iDriving Consortium – 101147004 Page 103 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Related Technical requirements TR-FUN-4.1-8 Table 23 - Red Light Violation Detector – Architectural Design and Description Red light violation detector: AI-Powered Behavioural analysis of objects of interest Description An intelligent vision-based module that monitors vehicle behaviour at intersections. It detects if a vehicle has crossed a virtual stop line during a red traffic light phase by combining real-time vehicle detection and tracking, traffic light data and rule-based violation detection logic. This component supports evidence generation for enforcement and real-time alerts for traffic management systems. Architectural Diagram and Subcomponent s (components View) Subcomponents: - Detector: Locates vehicles in each vide frame. - Tracker: assigns unique IDs to each vehicle and maintains its trajectory - Violation Engine: verifies virtual line crossing, evaluates whether the crossing occurred during a red traffic light phase, and triggers an alert. - Exporter: saves evidence (image + timestamp + ID + traffic light status) Expected TRL TRL 5-6 @iDriving Consortium – 101147004 Page 104 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Technologies The following technologies will be used for the implementation of the Component: - The software shall be implemented primarily in Python - SQLite, PostgreSQL, or MongoDB for export Deployment Premises On-premises service Involved Partners CERTH, TEKNIKER, SIMAVI, UNIV-EIFFEL Related Technical requirements TR-FUN-4.2-1 Table 24 - Improper Lane Usage Detector – Architectural Design and Description Improper line usage detector: AI-Powered Behavioural analysis of objects of interest Description This module detects improper lane usage by vehicles in urban segments. The system analyses the trajectory of each vehicle in realtime to determine whether it deviates from its assigned lane or enters restricted zones. It uses object detection, tracking, and spatial reasoning to identify violations. @iDriving Consortium – 101147004 Page 105 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Architectural Diagram and Subcomponents (components View) Subcomponents: - Detector: Locates vehicles in each vide frame. - Tracker: assigns unique IDs to each vehicle and maintains its trajectory - Violation Engine: validates vehicle trajectories based on the land zone. It compares the vehicle’s trajectory with the allowed lane zones. - Exporter: saves evidence (image + timestamp + ID + lane info) Expected TRL TRL 5-6 Technologies The following technologies will be used for the implementation of the component: • Python Deployment Premises On-premises Involved Partners CERTH, TEKNIKER, SIMAVI, UNIV-EIFFEL Related Technical requirements TR-FUN-4.2-2 Table 25 - Zebra Crossing Violation Detector – Architectural Design and Description Zebra crossing violation detector: AI-Powered Behavioural analysis of objects of interest @iDriving Consortium – 101147004 Page 112 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Involved Partners DREVEN Related Technical requirements TR-FUN-4.4-1/TR-FUN-4.4-2/TR-FUN-4.4-3 @iDriving Consortium – 101147004 Page 113 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Table 29 - AI-Based Real-Time Weather Alert System – Architectural Design and Description AI-Based Real-Time Weather Alert System: Smart Environmental Condition Monitoring for Proactive Road Safety Measures Description This system enables real-time and forecast-based environmental monitoring by integrating station observations and weather model predictions. Data is made accessible for specific geographic areas, and an automated process (scheduled via cron) checks if predefined hazard thresholds are met. The outcomes are processed by an AI engine (LLM), which generates alerts enriched with risk levels and weather-related data. These alerts are then served through an API, making them available to the iDriving platform for enhanced decision-making and road safety measures. Architectural Diagram and Subcomponents (components View) @iDriving Consortium – 101147004 Page 114 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Expected TRL TRL 6 Technologies • Python • Docker • Git • Django Rest Framework • Relational Database (e.g., PostgreSQL) • Open Source or Cloud-Based LLM (via API access) Deployment Premises DREVEN in-house servers or Cloud infrastructure Involved Partners DREVEN Related Technical requirements TR-FUN-4.4-4 @iDriving Consortium – 101147004 Page 115 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Table 30 - 3D-Smart Tool – Architectural Design and Description 3D-Smart Tool: AI-driven 3D Scene Generation for Safety & Maintenance Monitoring Description The tool aims to transform visual data from road environments into 3D representations, integrating it into the iDriving ecosystem for improved safety, maintenance insights and incident/progress monitoring. This involves acquiring data using UAVs, processing it using Neural Radiance Fields or Gaussian Splatting, and enhancing iDriving's ability to detect and assess road hazards, improving situational awareness and remote responses for road managers. Architectural Diagram and Subcomponents (components View) Expected TRL TRL 6 Technologies The following technologies will be used for the implementation of the 3D-Smart Tool: • 3D Gaussian Splatting • Neural Radiance Field @iDriving Consortium – 101147004 Page 116 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Deployment Premises - Involved Partners ACCELIGENCE, TEKNIKER, SIMAVI Related Technical requirements TR-FUN-4-5 @iDriving Consortium – 101147004 Page 117 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Table 31 - Dynamic Monitoring Platform – Architectural Design and Description Dynamic Monitoring Platform: Dynamic Update of Criteria Catalogue Through Continuous Learning Description The tool aims to analyse and predict the state of multimodal transport networks by integrating real-time and theoretical data. Using a generic and multimodal data model, it enables network monitoring, diagnosis, simulation, and forecasting based on key performance indicators like congestion, resilience, and speed. With the ability to interact with external tools via APIs and web services, it provides dynamic mapping, incident diagnosis, and performance visualization, ensuring enhanced safety and maintenance insights. Architectural Diagram and Subcomponents (components View) @iDriving Consortium – 101147004 Page 118 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Expected TRL TRL 6 Technologies The following technologies will be used for the implementation of the Dynamic Update of Criteria Catalogue Component: 1. Python Learning models Deployment Premises - Involved Partners - Related Technical requirements TR-FUN-5.1-1 TR-FUN-5.1-2, TR-FUN-5.1-3, Table 32 - SUMO: Digital Twin Powered Predictive Safety Measures and Warning Systems – Architectural Design and Description SUMO: Digital Twin Powered Predictive Safety Measures and Warning Systems Description The tool aims to simulate and analyse intermodal traffic systems, integrating road vehicles, public transport, and pedestrians for enhanced traffic management. By utilizing real-world traffic data, road topology, and SUMO add-ons like VANETT for telecommunications, it enables detailed traffic flow simulation, risk prediction, and hazard assessment. The platform supports real-time and large-scale simulations, requiring computational resources based on service levels, making it a valuable tool for optimizing urban mobility and safety. @iDriving Consortium – 101147004 Page 119 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Architectural Diagram and Subcomponents (components View) Expected TRL TRL 6 Technologies The following technologies will be used for the implementation of the SUMO: Digital Twin Component: 2. Sumo for traffic simulation of use case networks 3. Python for Safety assessment 4. Learning algorithms (Deep learning/ statistical methods) Geojson for safety alerts Deployment Premises - @iDriving Consortium – 101147004 Page 120 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Involved Partners TUC Related Technical requirements TR-FUN-5.3-* Table 33 - CARLA: Digital Twin Powered Predictive Safety Measures and Warning Systems – Architectural Design and Description CARLA: Digital Twin Powered Predictive Safety Measures and Warning Systems Description The tool aims to simulate and analyse traffic environments in 3D, providing insights for autonomous driving systems. By integrating real-world traffic data, road types, and obstacles, it enables risk prediction and hazard visualization. @iDriving Consortium – 101147004 Page 121 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Architectural Diagram and Subcomponents (components View) Expected TRL TRL 6 Technologies The following technologies will be used for the implementation of the CARLA: Digital Twin Component: • CARLA for simulation of monitored areas • Developments in Python for Safety assessment  Trajectory recognition and processing  Training Deep learning/ statistical models  Evaluation of the models Deployment Premises - @iDriving Consortium – 101147004 Page 128 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Architectural Diagram and Subcomponent s (components View) Expected TRL TRL6 Technologies The following technologies will be used for the implementation of the X Component: • Unity • C# • REST API • Photoshop • Blender 3D Deployment Premises Laptop (Windows), XR glasses Involved Partners INTRA, INFRA PLAN, ALP.LAB, AIM, DREVEN, THESSALONIKI Related Technical requirements TR-FUN-5.5-1, TR-FUN-5.5-2, TR-FUN-5.5-3 5.5 iDRIVING Process Views The following section presents component-level flow diagrams in the form of system sequence diagrams, contributed by all partners. Each diagram illustrates the runtime interaction between key elements of the respective component. A brief description accompanies each diagram to clarify its purpose and integration within the iDRIVING architecture. @iDriving Consortium – 101147004 Page 129 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies 5.5.1 Tekniker Dataspace Connector A modular solution that allows organizations to establish a single point of entry for accessing and exchanging data within a data space. It ensures interoperability in data sharing, fosters trust among participants, and guarantees data sovereignty throughout its entire lifecycle. Figure 11 - Tekniker dataspace connector sequence diagram Sequence diagram steps, presented in Figure 11: 1. The Data Owner defines their data catalog: a. Brief catalog description b. Catalog location (DataService) c. List of datasets: i. Dataset location (DataAddress) ii. Data type (Distribution) iii. Access policies (Offer) 2. The Data User requests the catalog 3. The Data User’s connector initiates the Catalog Protocol 4. The Data User receives and accepts the catalog's access policies (Contract Negotiation Protocol) a. Both parties can monitor the contract negotiation process 5. The Data User requests a dataset → the Transfer Process Protocol starts 6. Data transfer depends on the chosen method (PUSH or PULL): a. PULL: Processing tools access the data directly b. PUSH: Data is delivered to a defined location via the processing tools @iDriving Consortium – 101147004 Page 130 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies 5.5.2 Route Guidance tool Description of the sequence diagram: The process starts with sensors (cameras, GPS, road sensors, etc.) detecting incidents, congestion, or hazardous conditions. This data is analysed by the relative iDriving component, which extracts key traffic parameters. The results are communicated to the ITMS which then evaluates the situation and determines optimal traffic management strategies (based on the simulated scenarios), such as dynamic signal adjustments and route recommendations based on vehicle type. The system communicates with traffic signals to implement changes and provides guidance to road users (drivers, truck drivers, motorcyclists, cyclists) through the iDriving user communication channels (app, in-vehicle etc.). In critical situations, an Emergency Response System is also triggered for emergency vehicle drivers to ensure timely intervention. Figure 12 - Route Guidance tool sequence diagram Sequence diagram steps, presented in Figure 12: 1. Receive input: Data from various sources is collected and processed by the relative iDriving components. Useful information regarding actual or imminent traffic congestion is extracted and provided to the route guidance algorithm as input. 2. Historical data of travel demand, traffic congestion, traffic signal settings and other network-related information will be used for conducting SUMO microscopic simulation tests @iDriving Consortium – 101147004 Page 131 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies 3. Alerts and warnings about identified incidents (e.g. accidents, loopholes or infrastructure failure, fallen trees, extreme weather conditions) at specific location and time in the network (road, lane), which are provided by the relative iDriving tools, are given as input to the routing algorithm. The type, location and time of the incident are communicated. 4. All input data are introduced in the SUMO simulation environment of the respective traffic network (town/city) under consideration. Incident location is used to define the affected area. Traffic congestion levels, potentially hazardous weather conditions (e.g. wind, flooding, storm, infrastructure failure etc.) are used to estimate the value of a risk function for every link in the affected area per vehicle type. 5. Paths of all trips that are crossing the affected area are recalculated by the Route Guidance Tool’s algorithms, based on the current vehicle location and its destination. Alternative paths are decided so that the total risk level of all used paths is minimized among alternative options. 6. Alternative paths are communicated to the respective vehicles, through the relative iDriving communication components. 5.5.3 Signal Control Tool Figure 13 - Signal Control tool sequence diagram @iDriving Consortium – 101147004 Page 132 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Sequence diagram steps, presented in Figure 13: 1. Receive input: Data from various sources (e.g. cameras, GPS, and loop detectors) is collected and processed by the relative iDriving components. 2. Detection: Useful information regarding detected vehicles at intersections and/or queue length estimation at intersections per link, is provided to the signal control algorithm as input. 3. Signal Control Module: The module is introduced in the SUMO simulation environment with the appropriate historical data of travel demand, traffic congestion, traffic signal settings and other network-related information, to be used for conducting SUMO microscopic simulation tests. According to the number of vehicles waiting in the queues, traffic light splits are computed and broadcasted to the traffic lights. 4. Export: Finally, during the export, traffic light splits are stored in appropriate output files. 5.5.4 Mobile Application & In-Vehicle Application Figure 14 - Mobile Application & In-Vehicle Application sequence diagram Sequence diagram steps, presented in Figure 14: 1. Preparation: Mobile/in-vehicle application is deployed on the mobile device and running properly. @iDriving Consortium – 101147004 Page 133 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies 2. Connection: Mobile/in-vehicle application is connected to the internet and to the server. 3. Send/receive: Mobile/in-vehicle application sends/receives messages for different situations. 4. Server/Message bus: Server/Message bus sends/receives information. 5.5.5 Aerial Surveillance UAV System Deliver a reconfigurable UAV system equipped with different sensors and cameras to enable precise, efficient data collection from areas of interest on demand or in real time. This task aims to develop a UAV capable of autonomous navigation and real-time decision-making, particularly in complex environments such as maintenance operations and incident monitoring. By leveraging AI, advanced algorithms, and the unique capabilities of each UAV, the system will ensure a swift, accurate, and effective response to various road situations. Figure 15 - Aerial Surveillance UAV System sequence diagram @iDriving Consortium – 101147004 Page 134 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Sequence diagram steps, presented in Figure 15: 1. Prepare UAV Autonomous Flight: The UAV (Unmanned Aerial Vehicle) system is prepared for deployment, ensuring that the drone is operational, and all required parameters are set. 2. Drone Ready for Autonomous Flight: The UAV undergoes final system checks, including battery level, GPS calibration, and communication with the control platform. 3. Deploy UAV: The UAV takes off and begins patrolling the defined area. 4. The UAV Captures the required information such as maintenance operations and incident monitoring based on the selected sensors (RGB / Thermal cameras). 5. Image / Video Processing Using the Developed AI Algorithms 6. The platform receives information regarding any maintenance operations and incident monitoring. 5.5.6 Autonomous UAV Deployment for Area Coverage This component provides an AI-powered service for generating optimal flight paths for a single or multiple UAVs to cooperatively cover large offshore areas. It takes as input a geospatial area (in WGS84 format) and UAV parameters, and generates energy-aware, overlap-free trajectories. The paths account for obstacles, no-fly zones, and individual UAV constraints (e.g., battery, max range). The service is integrated into an end-to-end platform, where the frontend collects mission definitions and the backend processes them to deliver missionready flight plans in standard formats. Figure 16 - Autonomous UAV Deployment for Area Coverage sequence diagram @iDriving Consortium – 101147004 Page 135 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Sequence diagram steps, presented in Figure 16: 1. Input Source The input to the tool is a user-defined mission area, represented as a polygon in WGS84 coordinates, optionally including no-fly zones and static obstacles. Additionally, UAV specifications such as battery level, flight time, altitude limits, speed, and camera field-of-view (FoV) are required. 2. Data Processing The system parses and validates the geospatial input and UAV parameters. It then computes an optimal coverage path for either a single UAV or multiple UAVs based on the mission definition. Obstacles and no-fly zones are factored into the path planning algorithm to ensure legal and safe flight operations. In the case of multi-UAV operations, the system distributes sub-areas and computes coordinated coverage paths to ensure complete and efficient mission execution. 3. Output Source A standardized .JSON file is generated that includes: - The computed waypoints for each UAV (if multi-UAV) - Area and obstacles/No-Fly zones (if any) definitions - Mission parameters 5.5.7 Seat Belt and Cell Phone Detection Tool Figure 17 - Seat Belt and Cell Phone Detection Tool sequence diagram Sequence diagram steps, presented in Figure 17: 1. Input Source - The module receives data from street or drone cameras. - Input can be either individual frames or video streams. 2. Data Processing @iDriving Consortium – 101147004 Page 136 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies - A deep learning model detects traffic violations, including: o Drivers not wearing seatbelts o Drivers using mobile phones o Bikers not wearing helmets - Each frame is processed to detect relevant objects, with output including o Object class o Confidence score o Bounding box coordinates 3. Metadata Generation A JSON file is created containing metadata for each detected object. The system also counts the number of detected cars per frame and exports this information. 4. Output Generation The processed frames or video are forwarded to the iDriving UI. Detection results (objects and metadata) are also sent to TEKNIKER for further processing in the activity recognition task. 5.5.8 License Plate Detection Tool Figure 18 - License Plate Detection Tool sequence diagram Sequence diagram steps, presented in Figure 18: 1. Input Source The input source for this tool is the same as the detection tool—frames from street cameras or drone cameras. 2. Data Processing The tool processes the frame simultaneously with the detection module. It focuses on identifying vehicle license plates and extracting their numbers. Once a license plate is detected, it links this information to the @iDriving Consortium – 101147004 Page 137 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies corresponding offending vehicle (e.g., not wearing a seatbelt, using a phone, etc.). 3. Metadata Generation The output is a JSON file containing the license plate numbers of the offending vehicles, along with relevant metadata. 4. Output A JSON file is generated that includes the license plates of all identified offending vehicles in the processed frame. 5.5.9 Helmet Detection Tool Figure 19 - Helmet Detection Tool sequence diagram Sequence diagram steps, presented in Figure 19: 1. Input Source - The module receives data from street or drone cameras. - Input can be either individual frames or video streams. 2. Data Processing - A deep learning model detects traffic violation (bikers not wearing helmets) - Simultaneously we detect the license plate of each offender - We combine the information of the vehicle and the license plate 3. Metadata Generation - A JSON file is created containing metadata for each detected object. 4. Output The processed frames or video are forwarded to the iDriving UI. Detection results (objects and metadata) are also sent to TEKNIKER for further processing in the activity recognition task. @iDriving Consortium – 101147004 Page 144 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies 5.5.17 Pothole Detection and Severity Assessment Tool Figure 27 - Pothole Detection and Severity Assessment Tool sequence diagram Sequence diagram steps, presented in Figure 27: 1. Accept Input Sources - The module receives input from various sources, including drone cameras, fixed traffic cameras, and dashboard cameras in vehicles. 2. Select and Process Input - After selecting the appropriate source, the module feeds the data into a deep learning model designed to detect road defects such as potholes and cracks. 3. Generate Output - The model processes the input images and outputs annotated images, marking defect locations with bounding boxes. - Simultaneously, the same information is formatted as JSON text, which includes metadata like GPS coordinates (if provided by the original source). 4. Assess Defect Severity - The processed data is sent to the severity assessment module, where each defect’s severity is estimated. 5. Forward Results to UI - Finally, all results, along with severity assessments, are sent to subsequent modules and integrated into the project's UI, where they are presented. @iDriving Consortium – 101147004 Page 145 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies 5.5.18 Weather Prediction Tool (WRF and Data Assimilation) Figure 28 - Weather Prediction Tool sequence diagram Sequence diagram steps, presented in Figure 28: 1. Set Up Environment Deploy WRF and WRFDA within a Docker container for easy management and portability. 2. Data Preparation Gather meteorological data (e.g., GFS, observational data) required for initial conditions and boundary conditions. 3. Data Assimilation Use WRFDA to integrate real-time observational data into the model, improving initial conditions for better forecast accuracy. 4. Run Weather Prediction Execute WRF model using the prepared input data and assimilated fields for weather prediction. 5. Post-Processing Generate forecasts and relevant meteorological outputs (e.g., precipitation, temperature) for analysis and decision-making. 6. Verification Compare model outputs with real-time data to ensure forecast quality and accuracy. @iDriving Consortium – 101147004 Page 146 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies 5.5.19 AI-Based Real-Time Weather Alert System Figure 29 - AI-Based Real-Time Weather Alert System sequence diagram Sequence diagram steps, presented in Figure 29: 1. Forecast & Nowcast Retrieval Collect forecast data from external sources and generate current (nowcast) weather data. 2. Location-Based Filtering Extract and organize data for predefined discrete locations of interest (e.g., cities, road segments). 3. Hazard Identification Apply threshold-based rules to detect potential hazards based on forecast and real-time data. 4. AI-Driven Alert Generation If thresholds are exceeded, feed the event into an AI model (LLM) which creates a context-aware hazard alert. 5. Alert Delivery Alerts are exposed via a REST API and can be accessed in real time by the iDriving platform or other systems. @iDriving Consortium – 101147004 Page 147 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies 5.5.20 3D-SMART Tool Figure 30 - 3D-SMART Tool sequence diagram Sequence diagram steps, presented in Figure 30: 1. The tool proposes the UAV’s path based on the target area’s extent, ensuring that the scene will be captured from all possible viewpoints. 2. Raw data acquired from the UAV is extracted, processed, and formatted before being fed into the model. 3. The model receives the processed data, optimizes its parameters, and transforms the 2D data into a highly detailed 3D environment 4. Realistic 3D renders or gaussian splats in ply format can be extracted. 5.5.21 ClaireSITI Platform Figure 31 - ClaireSITI Platform sequence diagram Sequence diagram steps, presented in Figure 31: 1. Access data from use cases @iDriving Consortium – 101147004 Page 148 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies - Collect data from each use cases (UAVs, weather stations, AI camera outputs) based on SCC requirements 2. Process to the computation of KPI based on new data collected 3. Update SCC formula/parameters through continuous learning of KPI 4. store KPI send updates for all partners or road managers through frontend dashboard or geojson 5.5.22 SUMO (Simulation of Urban Mobility) Provide alerts and safety measures based on a digital twin built using the SUMO simulator. The tool is designed to simulate various transport scenarios using historical data and real-time field data (UAVs, cameras, sensors) to anticipate risks for different road users and deliver alerts or safety measures to users and network managers. Figure 32 - SUMO sequence diagram Sequence diagram steps, presented in Figure 32: 1. Access data from monitored areas a. Collect data from edge sensors (UAVs, weather stations, AI camera outputs) to adapt traffic scenarios based on real-time data. 2. Access historical data for traffic calibration a. Use historical data to determine initial traffic demand across the network, including areas outside monitored zones. 3. Simulate traffic scenarios in SUMO a. Perform a simulation of road users in SUMO, integrating both historical demand and real-time data. SUMO also integrates @iDriving Consortium – 101147004 Page 149 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies communication tools to simulate various scenarios according to road user compliances to IDriving systems and communication delays 4. Process safety assessment on simulated data a. Extract safety features from simulated data—such as user trajectories and behaviours—to compute safety measures or alerts. 5. Provide geolocated alerts and measures a. Send safety measures, alerts, or risk assessments across the network based on infrastructure type, formatted as GeoJSON. 5.5.23 CARLA (Autonomous Driving Simulation Platform) Provide alerts and safety measures based on a digital twin built using CARLA and SUMO simulators to ensure a 3D representation on monitored area. The tool is designed to simulate various transport scenarios using historical data and real-time field data (UAVs, cameras, sensors) to anticipate risks for multiple road users (car, pedestrian, cyclist) and deliver alerts or safety measures to users and network managers. Figure 33 - CARLA sequence diagram Sequence diagram steps, presented in Figure 33: 1. Access data from monitored areas a. Get data from on edge sensors (UAV, weather station, IA camera outputs) to detect incoming road user on the monitored area 2. Access historical data for traffic calibration a. Historical data provides initial demands of the whole network outside monitored areas 3. Simulate in 3D road user behaviours on CARLA and SUMO @iDriving Consortium – 101147004 Page 150 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies a. Realise a 3D simulation of the monitored areas in use cases based on incoming road users and historical demands. Various road user characteristics (aggressive, risky) can be provided for simulation. 4. Process safety assessment on simulated data a. Extract safety features from simulated data – user trajectories, behaviours – to compute safety measures or alerts 5. Provide alerts and measures a. Interact with roads users on site with safety measures by providing potential risks 5.5.24 AI-Optimized Maintenance through Digital Twin Consists of: • Module 1.1: Dynamic Risk Assessment • Module 1.2: Health and Logistics Management • Module 1.3: Maintenance Scheduling The system relies on a shared infrastructure and data pipelines to support decisionmaking in near real-time. Global Component Sequence Overview Figure 34 - AI-Optimized Maintenance through Digital Twin - Global Component sequence diagram Sequence diagram steps, presented in Figure 34 Data flow and Module interactions: @iDriving Consortium – 101147004 Page 151 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies 1. Sensor Network collects real-time data like condition, weather, and traffic, and sends it to Module 1.2 (Health & Logistics Management). 2. Module 1.2 performs data processing (ETL) and computes health indices (like RCI and HI), then displays this info on the Dashboard. 3. Module 1.2 also sends processed health and vulnerability data to Module 1.1 (Dynamic Risk Assessment), which: • Calculates risks using methods like FMEA. • Runs simulations (e.g., “what if" scenarios). • Outputs result to the Dashboard (e.g., risk maps). 4. Module 1.1 passes risk and scenario data to Module 1.3 (Maintenance Scheduling), while Module 1.2 also provides logistics info. 5. Module 1.3 combines inputs to generate shortand long-term maintenance plans, cost estimates, and priority lists—also visualized on the Dashboard. Module 1.1 Risk Assessment Figure 35 - AI-Optimized Maintenance through Digital Twin – Module 1.1 Risk Assessment sequence diagram @iDriving Consortium – 101147004 Page 152 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies Sequence diagram steps, presented in Figure 35: 1. Receive data from Module 1.2: Inputs include the Health Index, traffic, road, and vulnerability information. 2. Compute Risk using FMEA Methodology: Module 1.1 uses Failure Modes and Effects Analysis (FMEA) to assess potential risks and their impact. 3. Run Scenario Simulation – What-if Engine: It simulates various hypothetical conditions to test system resilience and response under different stressors. 4. Generate Risk Metrics: Outputs include quantifiable risk values, rankings, or categories that can inform decision-making. 5. Expose REST API to Dashboard: Module 1.1 shares results via an API, enabling integration and real-time access. 6. Dashboard displays risk maps and simulations: The frontend/dashboard visualizes these insights for users—e.g., risk heatmaps, simulation outcomes. Module 1.2 Health & Logistics Management Figure 36 - AI-Optimized Maintenance through Digital Twin – Module 1.2 Health & Logistics Management sequence diagram Sequence diagram steps, presented in Figure 36: 1. Get External Forecasts: Module 1.2 receives weather and traffic forecasts from external APIs. 2. Step 2: Stream Sensor Data: It simultaneously ingests real-time condition data from on-field sensors (e.g., pavement RCI). @iDriving Consortium – 101147004 Page 153 of 157 Intelligent & Digital Roadway Infrastructure for Vehicles Integrated with Next-Gen Technologies 3. ETL Processing: Extraction of raw data, Transformation (cleaning, normalization), Loading and integration into its system 4. Predictive Modelling: It runs machine learning models to compute a Health Index for each road segment. 5. Send to Dashboard: The final output—road condition forecasts and Health Index—is pushed to the dashboard for end-user visualization. Module 1.3 Maintenance Scheduling Figure 37 - AI-Optimized Maintenance through Digital Twin - Module 1.3 Maintenance Scheduling sequence diagram Sequence diagram steps, presented in Figure 37: 1. Step 1: Input from Module 1.1 Receives risk levels and scenario outputs that indicate potential failure zones and risk-critical areas. 2. Step 2: Input from Module 1.2 Gets logistics and condition metrics, including road degradation, accessibility, and available resources. 3. Step 3: Optimization Engine