Continuums Of Game NETs: swarm intelligence as information processing
IoTs devices currently lack the ability to self-organizate and learn collaboratively i.e., they cannot autonomously recognize and respond in respect to the availability of data and computing resources that are dynamically changing.
CoGNETs aims to develop a Middleware Framework that will empower IoT, Edge, and Cloud devices to autonomously organize dynamic IoT-to-Cloud swarm continuums for optimal data processing and seamless service provisioning.
What is needed is to approach the problem with beyond-state-of-art AI designs that can end up deploying intelligence as integral part of each device (i.e. as instinctive behaviour) to support every activity, process, and decision-making, ranging from self-organization of on-board datasets and hardware resources, to collaborative-learning in mesh local-Clouds. The idea is to enable devices to self-realize their heterogeneity and self-decide how to form dynamic IoT-to-Cloud swarm continuums to respond to common AI tasks and significant elements of cognitive computing in an automated, secure and energy-efficient manner. At its core, our middleware will integrate a new decentralized federated multi-context broker architecture enriched with game-intelligent agents, collaborative federated learning, and end-to-end security mechanisms (including RISC-V hardware security and AI acceleration) to effectively manage dynamic data and resources across devices in different domains, using open-source programming models and APIs to create new orchestration, optimization, etc. routines without vendor lock-in. Validations will be conducted via 3x (per-vertical) plus 1x (cross-vertical) demos to support key sectors of manufacturing, mobility and health, and to examine how CoGNETs can ensure economic, business and societal resilience, and how it can extend Europe’s digital sovereignty, strategic autonomy and industrial dominance in the next wave of technology evolution.
Security and Privacy are fundamental aspects in self-governing swarms. SBA Research will lead the end-to-end security task and focus on Application-level and AI-level security. This will include strategies to protect the federated learning mechanisms and trained models.
CoGNETs project aims to develop an intelligent IoT-to-Cloud computing continuum using dynamic swarms of autonomous devices. By integrating game theory, collaborative federated learning, and end-to-end security, it seeks to optimize the balance between computing, energy, and security. This will enable efficient, secure, and responsive AI-driven services across various sectors, including manufacturing, mobility, and health.
Project Lead: CERTH
Sustainable Development Goals
Helping the EU Society to achieve sustainability goals, by offering a fully automated Middleware solution that inherently minimizes the energy consumption of its underlying devices during operation, thereby slowing down the impacts of climate change and helps the EU Society embracing environmentally conscious behaviours via automation technology.
Support the Green Deal and the Green Transition of the EU Data & AI
Economy: Contribution to the sustainable development and the high rank that Europe holds in Environmental standards and green transition by directly supporting GHG/CO2 footprint reduction through (1) optimization of Energy in joint and equal terms to Computing and Security performances (i.e. joint Computing-Energy-Security KPIs as Game optimization utilities), (2) push of complex AI layer processing to the Cloud while keeping lightweight AI layers at the IoT/Edge (i.e. via Collaborative Federated Learning), (3) development of environmentally friendly applications, like collaborative manufacturing robots and BEV vehicle fleet strategies with jointly optimized Computing-Energy-Security service performances, thereby inspiration of end-users to adopt carbon reduction practices.