WatchEDGE studies an advanced edge-computing architecture supporting AI-based applications distributed on geographically-distant sites. The objective of WatchEDGE project is to create a geographically-distributed infrastructure to support surveillance applications for the rural environment, in order to defend it from natural threats such as wild animals, wild fires, etc. Each site (or “island”) – interconnected by SD-WAN – is equipped with edge-computing infrastructure that can be provided by fleets of flying drones (FANET), and smart radars and fixed or flying cameras. environmental surveillance is nowadays a typical task for Artificial Intelligence (AI), which in this case will entail Vision Computing (VC) functions for identification and classification.
The project – orchestrated to maximize data processing at the edge – works towards a use case of environmental surveillance for smart agriculture and wildlife protection, relying on AI-based image processing.
WatchEDGE is part of Spoke 8 – Intelligent and Autonomous Systems
Project PI: Guido Maier
Work package 1 activity is dedicated to "Architecture and Requirements", and focused on the preparation of the first deliverable. The deliverable identifies the application scenario of the WatchEDGE project, the requirements, and the main features of the architecture. The project was initially participated by four partners (PoliMI [leader], UniCatania, CNIT, Italtel). On March 2024, four new partners (UniPisa, UniBicocca, Nextworks, Sensor ID) joined the project by means of the cascade calls.
Work Package 2 focuses on developing orchestration algorithms for the WatchEDGE architecture, integrating wildlife monitoring, AI-based computer vision, UAV control, and SD-WAN orchestration across the far-edge–edge–cloud continuum. The team designed a cloud–edge orchestration framework supporting multi-site surveillance, where SD-WAN overlay tunnels connect distributed “islands” to ensure secure and efficient routing. AI models are deployed directly at the edge using smart cameras provided with embedded computing capability, which combines with a newly built wildlife dataset for model training, validation, and visualization through heatmaps. Radar equipment is also used in combination with optical sensors to enhance animal detection accuracy. Additionally, a resource-management framework for FANET inference placement was developed, optimizing virtual-function deployment on UAVs under delay and energy constraints. These developments collectively enable dynamic orchestration, efficient routing, and real-time environmental monitoring within WatchEDGE.
Work Package 3 focused on designing and validating the WatchEDGE Proof of Concept (PoC) in collaboration with the San Rossore Park in Pisa. The main goal was to demonstrate the feasibility and performance of the architecture, orchestration, and algorithms developed in previous work packages. The primary use case addresses wildlife monitoring and image processing using far-edge trap cameras and drones equipped with YOLOv8n for real-time and offline animal detection. These systems created annotated datasets for species recognition and behavioral analysis. Additional experiments included vegetation health assessment with multispectral drones and a radar system for motion detection of wild animals. The WatchEDGE infrastructure integrates edge computing, SD-WAN, and private 5G (Nomad) connectivity, managed by the WatchEDGE Orchestrator. A dashboard prototype visualizes distributed devices, events, and heatmaps of detected wildlife or vegetation data, while allowing operators to tag and catalogue scenes—enhancing datasets for ongoing environmental monitoring.
They are contained in the deliverables of the project:
D.1.2. FINAL SYSTEM ARCHITECTURE AND REQUIREMENT DESCRIPTION
The document integrates:
Both the physical architecture and the orchestration architecture are original, designed specifically for WatchEDGE through the joint work of all partners. The system introduces strong innovation by creating an integrated platform tailored to environmental monitoring for rural communities. In the architecture, sensors on the ground and on drones collect data, local computing units process information, and multiple customer sites interconnect to share resources. A multi-level orchestration framework coordinates applications, sensors, drones, and network connectivity so the system can operate efficiently, autonomously, and continuously.
The deliverable also outlines the essential requirements for sensors, drones, computing elements, and the overall control system. These architecture and definitions were defined in a preliminary form in Deliverable D.1.1.
D.2.1. WATCHEDGE PLATFORM AND ALGORITHMS:EARLY SOFTWARE RELEASE AND REPOERT
The main results achieved in this deliverable are the following::
This deliverable unrolls the implementation details of the WatchEDGE system. It summarizes the developments carried out in WP2 during the project, focusing on SD-WAN connectivity, ground sensors, UAV/FANET management, and AI-based orchestration. The document reports how partners jointly designed innovative mechanisms to support reliable connectivity across remote customer sites, developed radar and camera-based detection pipelines, and defined methods for using UAVs both for monitoring and for distributed AI tasks. It also presents the orchestration framework that coordinates services, connectivity, and edge resources. Overall, this deliverable shows WatchEDGE’s AI-enabled system capable of supporting environmental monitoring, wildlife detection, and other rural services through a combination of sensors, UAVs, edge computing, and intelligent orchestration.
D3.1: REPORT ON VALIDATION AND PERFORMANCE ASSESSMENT (WP3)
Key Contributions of Deliverable D3.1:
Deliverable D3.1 reports the validation and performance assessment of the WatchEDGE platform across its major use cases and technological components. Over 22 months, partners conducted extensive experimental work to verify the feasibility, performance, and practical impact of the algorithms, architectures, and orchestration mechanisms developed in WP1–WP2. The document covers field experiments for wildlife monitoring, vegetation analysis, radar-based detection, UAV-based sensing and computing, front-hauling and back-hauling connectivity, SD-WAN inter-island networking, datasets and AI training pipelines, and distributed FaaS orchestration. A large on-field Proof of Concept was deployed in San Rossore Park, integrating cameras, drones, edge/far-edge compute, radar, AI models, SD-WAN, and orchestration tools. Findings confirm that WatchEDGE can operate autonomously in rural environments, coordinate heterogeneous resources, and support AI-driven environmental monitoring across multiple sites. The work provides practical insights that will be fed back into WP1–WP2 for architectural and algorithmic refinement.
WatchEDGE has carried out extensive dissemination activities targeting scientific communities, industry stakeholders, and the general public. Project results were also presented during several RESTART plenary meetings in Bari, Napoli, and Bologna, ensuring continuous interaction with the broader RESTART community.
The experimental phase of the WatchEDGE project officially started within the RESTART program. A joint press release was published in June 2025 by Politecnico di Milano, the University of Pisa, the RESTART consortium, and the Parco di San Rossore.
The press release, titled “WatchEDGE al via la sperimentazione; inelligenza artificale al servizio della natura per monitorare la fauna selvatica”, announced the deployment of the experimental use case at the Parco di San Rossore (Pisa). The pilot focuses on wildlife monitoring, targeting species such as wild boars, fallow deer, and wolves. The initiative combines artificial intelligence–based image analysis, next-generation sensing technologies, and innovative connectivity solutions to enable scalable and sustainable environmental monitoring in protected natural areas.
Newspaper Interviews and Media Coverage
WatchEDGE received wide media attention through several newspaper interviews and articles. Below, a selection of representative contributions is reported.
In March 2025, La Sicilia published an interview with Prof. Giovanni Schembra (University of Catania), highlighting the role of the Aerial Monitoring Laboratory of the Department of Electrical, Electronic and Computer Engineering. The article described how UAVs developed by the laboratory are being used to monitor the Parco di San Rossore in Pisa.
In May 2025, Innovation City featured an interview with Antonino Albanese (Italtel), focusing on the societal and environmental impact of WatchEDGE. The article emphasized the project’s objective of providing farmers, park managers, and rural operators with advanced tools to monitor critical phenomena, ranging from wildfires to plant diseases. WatchEDGE is developing a technological infrastructure capable of supporting AI-based image surveillance applications, with potential deployment in agriculture, forestry, and environmental protection. Particular attention was given to wildlife management, an increasingly relevant challenge in rural areas where animal presence can cause agricultural damage and facilitate the spread of diseases among livestock.
In November 2025, Il Tirreno published an interview with Prof. Stefano Giordano (University of Pisa), further discussing the project’s deployment in Tuscany and its role in combining edge computing, AI, and advanced networking technologies for environmental monitoring.
Public Workshops
WatchEDGE organized several public and scientific workshops aimed at engaging researchers, practitioners, and end users.
As part of its dissemination activities, WatchEDGE organized a public workshop titled “WatchEDGE Architecture and Requirements” in September 2023 at the University of Catania. The workshop presented key technical developments, including open-source orchestrators and simulation tools, object detection using open computer vision frameworks, supervised-learning platforms for aerial wildlife recognition, federated learning approaches for computational load balancing in FANETs, and the initial implementation of smart radar sensors. The event gathered researchers and practitioners to discuss ongoing progress and practical challenges.
In February 2024, WatchEDGE organized a public dissemination workshop titled “AI and Edge Continuum for Faunistic Monitoring”, held at Parco Migliarino, San Rossore e Massaciuccoli (Sala Gronchi, Pisa). The workshop presented key aspects of the project, including infrastructure design, orchestration architecture, far-edge and FANET hardware implementation, radar system development, and distributed learning approaches for edge-based machine learning. Additional sessions addressed computer vision algorithms for wildlife monitoring and zero-touch training solutions based on FANETs. The event brought together researchers and practitioners to discuss technical progress, experimental deployment, and future field trial activities within the WatchEDGE ecosystem.
From March 3 to 5 2025, the WatchEDGE plenary meeting was held at the San Rossore Park in Pisa. The program combined field experimentation and technical discussions among partners, with the first and last days dedicated to on-site trials and the central day focused on a public workshop held in Sala Gronchi. The workshop addressed both technical and non-specialist audiences, including park authorities and staff, and covered topics such as drone-based animal identification, wildlife census methodologies, proof-of-concept planning, and computer vision techniques for animal monitoring. Approximately 40 participants attended. A summary of the experimental results was later presented at IEEE WCNC in Milan at the RESTART booth.
From June 30 to July 1, 2025, WatchEDGE participated in the RESTART Plenary Meeting held in Naples. During the event, the project showcased its technologies through a dedicated WatchEDGE booth, where partners presented a live demonstration of the platform.
From July 14 to 16 2025, the 13th WatchEDGE plenary meeting was held at the San Rossore Park in Pisa. The program combined field experimentation activities with technical discussions among partners. The central day featured the public workshop “New Challenges for Wildlife Monitoring,” held at Officine Garibaldi in Pisa and open to both experts and a broader audience. The workshop addressed topics such as AI-driven wildlife monitoring, rural ecosystem impact, industrial development perspectives, and drone-based animal identification techniques.
On September 21 2025, WatchEDGE delivered a public educational lecture at the Festival Internazionale dell’Ingegneria hosted by Politecnico di Milano. The session introduced the project and presented topics including wildlife monitoring, network architecture design, 3D reconstruction, aerial environmental monitoring, and a live demonstration of animal classification technologies.
Project partners also contributed to the organization of major international events, including the 3rd and 4th International Workshops on Edge Network Softwarization (ENS 2024 and ENS 2025), co-located with IEEE NetSoft.
In the context of the IEEE International Conference on High Performance Switching and Routing (HPSR 2024), held in Pisa, Italy, on July 22–24, WatchEDGE co-sponsored and organized several dissemination activities open to conference participants. These included the workshop “Data-Plane Programmability and Edge Network Acceleration in the AI Era (NetAccel-AI 2024)” and the panel “Networked AI for Environmental Monitoring of Rural Areas,” which gathered academic and industry experts in different fields (information engineering, biology, zoology) to discuss emerging challenges and solutions. WatchEDGE also contributed invited talks such as “The Role of SD-WAN in Edge Network Softwarization.” In addition, the project participated in the RESTART “Future Vision Workshop” at Politecnico di Milano, further strengthening its scientific visibility and outreach.
WatchEDGE actively participated in the final RESTART Meeting “Shaping Horizons In Future Telecommunications”, Rome, January 19-21, 2026. During the demo sessions, our project successfully presented 7 demos at the WathcEDGE booth, which was extensively visited by delegates on both days of the conference. The videos of the demo are available on our Youtube channel (see below the “Videos and multimedia content” section).
Videos and Multimedia Content
In addition to press releases, interviews, and public events, WatchEDGE dissemination activities also include a set of dedicated video materials published on the official RESTART YouTube channel. These videos present the project from different perspectives, covering its vision, technical contributions, and experimental demonstrations.
A first video features the PI of WatchEDGE prof Guido Maier, who introduces the project objectives, the underlying technological approach, and the relevance of edge AI and SD-WAN solutions for environmental monitoring use cases. The interview highlights the role of WatchEDGE and its contribution to advancing AI-driven monitoring infrastructures.
Video link: https://youtu.be/R7wErUj34G8?si=rc4Gy0US6Q64r-kF
A second video focuses on the perspective of a young researcher involved in the project. This contribution provides insights into the different aspect of research activities in WatchEDGE, the experimental setup, and the challenges of WatchEDGE and solutions in real-world environments. The video also reflects the educational and training impact of WatchEDGE on young researchers.
Video link: https://youtu.be/kcXIyDyUy-k?si=20_6Txzdavh12Fk5
Then, a dedicated demo video presents the WatchEDGE platform in operation. The video showcases the end-to-end system, including data acquisition, AI-based image processing, federated learning workflows, and the interaction between edge nodes and the SD-WAN infrastructure. This demonstration illustrates how the proposed architecture supports scalable and efficient wildlife monitoring under realistic network and resource constraints.
Video link: https://www.youtube.com/watch?v=q0ncIza8ro4
In addition, WatchEDGE created a dedicated YouTube channel to host demo videos that were presented during RESTART events but were not published on the official RESTART channel. The channel serves as a repository for technical demonstrations and project updates, providing additional visibility to ongoing developments. Current videos include demonstrations of radar-based animal ranging and speed detection, decentralized deployment scenarios, monitoring dashboards, SD-WAN integration, etc.. These materials offer a practical view of the WatchEDGE platform and complement the project’s official dissemination outputs by showcasing real experimental setups and technical implementations.
Channel link: https://youtube.com/@watchedge-restart?si=IZ7oEnvkDn048x7I
These video materials complement the project’s written dissemination outputs and provide an accessible overview of WatchEDGE for both technical audiences and the general public.
Web Announcements and Broadcast Media
Project updates and announcements were published through the official communication channels of several partner institutions and stakeholders, including: RESTART Program, DEIB – Politecnico di Milano, Italtel, Parco di San Rossore and University of Milano-Bicocca.
In addition, WatchEDGE activities were widely shared through various social media platforms and online channels. WatchEDGE was featured in several broadcast and audio media outlets.
In December 2025, the project was presented on TG Leonardo (RAI 3), Italian national television, highlighting the role of AI and edge computing in environmental monitoring.
In May 2025, the biologist Dr. Marco Del Frate, co-operator of WatchEDGE and organizer of experimental activities in Parco di San Rossore, was interviewed on Vitamina G, a program broadcast by Radio Utopia.
In June 2025, Prof. Guido Maier (Politecnico di Milano) participated in an episode of the Tutto Connesso podcast, available on Spotify, discussing the project’s technological vision and societal impact.
Audio link: https://www.radioutopia.biz/
June 20, 2025, Podcast Tuttoconnesso, Guido Maier interview
Podcast link: https://open.spotify.com/episode/0gPQTCsjt1QhBrjyKVdB3U
Fair-like Events
The WatchEDGE project participated in Maker Faire Rome 2025. On October 17, Prof. Stefano Giordano (University of Pisa) presented the project, and a dedicated booth showcased WatchEDGE throughout the event. The participation was sponsored by the start-up Sensor-ID, whose team, together with young researchers from Politecnico di Milano and the University of Pisa, demonstrated the platform to visitors. Thanks to the booth and demonstrations developed by all project partners, WatchEDGE received the “Maker of Merit” award.
The project was also showcased at EXPO 2025 in Osaka, Japan, in May 2025, and participated in the Consumer Electronics Show (CES) in Las Vegas, USA, in January 2026, further expanding its international visibility.
The platform supports wildlife detection coordinating distributed AI services across geographically dispersed sites, but it can also be employed for smart agriculture, and early wildfire identification. Its architecture is inherently extensible and can be easily adapted to additional verticals such as large-scale IoT deployments, Industry 4.0, and Smart Cities.
In essence, WatchEDGE offers a unified and adaptable infrastructure that brings artificial intelligence to the edge, enabling real-time insights and autonomous decision-making across complex, multi-site environments.
The adoption of an automated computing/communications/sensing architecture and service automation platform can fulfill different use-cases leveraging on:
Each of the industrial partners involved (3 over 8) has a clear scope and mission within the project, which makes their expertise and contributions well integrated. Contributions of the industrial partners can be summarized as follows:
Cascade calls partners:
Community of workers and operators of the rural environment, such as farmers, national-park managers, etc., needs to protect their properties against natural threats, such as wild animals, crop parasites and diseases, raging fires. AI technology is the best allied to effectively control these threats, provided that the community can rely on a powerful and efficient platform able to support AI with enough computing and networking resources.
WatchEDGE studies an advanced edge-computing infrastructure distributed over several geographically-distant sites. It allows raw data to be processed at the edge, while computing functions and trained models are efficiently moved across, thanks to the WatchEDGE SDN-control and orchestration system. Each site is equipped with local edge-computing and networking resources, including fleets of flying drones (FANETs), while inter-site connectivity is provided by SD-WAN.
As of November 2025, WatchEDGE partners have produced more than sixty scientific contributions, including journal publications, conference papers, invited talks, and tutorials.
1. Publications
2. Joint Publications
3. Talks/Dissemination events
4. Demo/PoC
5. Patents/Innovations
6. Open source contributions
Accomplished 100%
Accomplished 100%
Accomplished 100%
Accomplished 100%
Accomplished 100%
Expected 8
Accomplished 11
Expected 12
Accomplished 35
Expected 4
Accomplished 8
Expected 2
Accomplished 2
Researchers involved: about 140 person-month (30 univ. permanent staff + 50 univ. RTD/PHD + 60 industry)
Collaboration Proposals:
- WathcEDGE is looking for possible collaboration with end users of the WatchEDGE solution, such as: wildlife parks, agricoltural institutions, farmers, breeders, etc.
- WatchEDGE is contributing to the following RESTART Grand Challenges: Challenge 0 – Create a vision of future evolution of telecommunications ecosystem in Italy and internationally; Challenge 7 – Digitalize the environment for a sustainable world; Challenge 10 – Make the network a platform for programming and running applications; Challenge 12 – Make artificial intelligence distributed and networked; Challenge 19 – Create a community for RESTART open source software
- The WatchEDGE proposal was endorsed by ASSOProvider by a letter of interest
- Within RESTART program, WatchEDGE is seeking co-operation with the structural projects NETWIN, SUPER and COHERENT and with the focused project LEGGERO
For any proposal of collaboration within the project please contact the project PI.
WatchEDGE News
| File | File size | Downloads |
|---|---|---|
F1 webpage WATCHEDGE images
|
9 MB | 7 |