WatchEDGE studies an advanced edge-computing architecture supporting AI-based applications distributed on geographically-distant sites. 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.
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
[M1-M9] In the first nine months of RESTART, most of WatchEDGE activity covered “Architecture and Requirements” and focused on the preparation of the first deliverable, that will identify the application scenario of the WatchEDGE project, the requirements, and the main features of the architecture.
Partners also started the activities on “Orchestration Algorithms and Methods” and disseminated the project goals internally and externally.
Procedures to buy experimental equipment are in progress.
[M9-today] WatchEDGE gave an important contribution to the definition of the overall platform architecture. Specifically, the main contribution has regarded the role of FANETs for both monitoring customer sites and providing networking, storage and computing on demand.
[M1-M9]
- Orchestrators for WatchEDGE: Open-Source Implementation and Simulation
- Object Detection with Open Computer Vision
- A Supervised-Learning Platform for Automatic Wild-Animal Aerial Recognition
- Optimizing Computational Load Balancing in FANETs for Wild-Animal Aerial Recognition: A Federated Learning Approach
- WatchEDGE Smart Radar Sensors: Initial Phase of Hardware Implementation
- Resource-Aware Model Aggregation for Federated Learning Training Rounds in SD-WANs
In addition, four technical papers were produced:
- “From MPLS to SD-WAN to ensure QoS and QoE in cloud-based applications”
- “Performance Characterization and Profiling of Chained CPU-bound Virtual Network Functions”
- “OSCAR: a Contention Window Optimization approach using Deep Reinforcement Learning”
- “MANTRA: an Edge-Computing Framework based on Multi-Armed Bandit for Latency- and Energy-aware Job Offloading in Vehicular Networks”
Detailed list of papers available on RESTART Mission 7 repository.
[M9-today]
- Main Scientific outcomes: definition of the role of the FANET Orchestrator for decision making regarding flight control (takeoff and landing) and virtual function placement on MEC-UAVs belonging to the FANET.
- Main Industrial/Exploitable outcomes: management and orchestration of a storm of UAVs organized in FANET.
- Publications (50%)
- Joint Publications (30%)
- Demo/PoC (5%)
- System Architecture definition (80%)
- Orchestration and Management Architecture (30%)
- Implementation (5%)
Project PI: Guido Maier
Collaboration Proposals:
WatchEDGE is currently looking for new partners, as four tasks of the project are currently included in the RESTART Cascade Calls related to SPOKE 8 – INTELLIGENT AND AUTONOMOUS SYSTEMS, coordinated by Università degli Studi di Roma Tor Vergata. Check the Call at the following link.
For any proposal of collaboration within the project please contact the project PI.
WatchEDGE News
