Edge AI Reference Architecture
Abstract
The edge AI reference architecture requires a technical system overview that provides a generic approach applicable to multiple industrial sectors, rather than one developed for a single industry (industrial-sector agnostic). A layered architecture that aligns with the edge AI technology stack should be considered. The layers should include hardware, hardware accelerators, interfaces, edge AI frameworks, software, edge AI models, algorithms, data, and applications. It can be connected to an IoT architecture, which includes layers such as sensors, connectivity, edge processing, communication, memory, storage, and edge computing. The three-dimensional architecture should consist of, in the second dimension, edge AI system quality properties (related to dependability, security, reliability, maintainability, and trustworthiness), and, in the third dimension, performance properties describing the edge AI system. The edge AI architecture can then be unfolded for each use case, and all three sides of the architecture can be mapped to the elements or components of edge AI systems, including hardware, edge AI frameworks, and edge AI models.
Full text
Copyright © 2025 1 European Conference on EDGE AI Technologies and Applications - EEAI 20-22 October 2025, Naples, Italy The intersection of imagination and execution, where edge AI learns to create, reason, and act.
Copyright © 2025 European Conference on EDGE AI Technologies and Applications - EEAI Milan, Ita20-22 October 2025 Naples, Italy 2 Copyright © 2025 Ovidiu Vermesan, SINTEF AS, Norway Edge AI Reference Architecture
Copyright © 2025 Presentation Outline •Reference Edge AI System Architecture •Reference System Architecture Role •Edge AI Technology Stack •Edge AI Layers and Planes 3
Copyright © 2025 Reference Edge AI System Architecture 4 •Areference edge AI architecture is a standardised, applicationagnostic model or framework that defines the fundamental components, their roles, and how to deploy and manage edge AI system workloads. •It serves as a blueprint to guide the creation, design, and build of concrete and specific, real-world edge AI systems, ensuring the quality properties and functional requirements for the edge AI system.
Copyright © 2025 Reference Edge AI System Architecture Role 5 •The reference edge AI system architecture features multiple views, each defined at varying levels of detail and abstraction. •Provides a common language/vocabulary for the different stakeholders involved in the area, assures the consistency of implementation of the edge AI technology stack to solve the technical challenges, and strengthens the verification, validation, testing, and benchmarking of solutions against a reference framework. •Supports the commitment to use common standards, specifications, and patterns for implementing the different use cases when applying edge AI techniques and methods in various applications for industry domains.
Copyright © 2025 Reference Edge AI System Architecture Role 6 •The reference edge AI system architecture aims to offer a high-level pattern/template solution for edge AI-based applications across the industrial domains, supporting a common vocabulary with which to discuss implementations in various use cases, analyse commonalities, and identify specificities and gaps. •The use of the reference edge AI system architecture serves as the basis for providing a methodology and set of practices and templates that generalise a set of previous solutions used in industrial domains. •The solutions are then applied to various use cases, generalised, and verified, validated, tested, and benchmarked for use in different industrial domains.
Copyright © 2025 Reference Edge AI System Architecture Functions 7 oSupports HW/SW partitioning and, at a lower level, enables the interactions of procedures (or methods) within an edge AI-based embedded IoT/IIoT, agentic AI, and generative AI system designated to perform a specific task. oAllows tracing the data/knowledge flows and streams between the various architectural layers and identifying the use of common data/knowledge formats as an integration method to avoid every connector/interface having to convert data to/from the different formats of every component, as well as the use of interfaces and APIs between components. oIt can be applied to edge AI distributed systems that share the computing functions and act as a coherent system to the overall application. oThe implementation, including agentic edge AI, allows the use of autonomous components that interact and collaborate to implement intelligent functions at the system level. HW/SW partitioning Distributed systems Data/knowledge flows
Copyright © 2025 Reference Edge AI System Architecture Use 8 External Heterogeneity Internal heterogeneity is defined as the ability of the system to accommodate numerous types of components and interfaces (using the same or different programming languages), as well as other algorithms/software modules from various sources. External heterogeneity is defined as the ability of the edge AI system to adjust to and leverage various HW/SW platforms, network protocols, operating systems, middleware infrastructures, and algorithms to meet the edge AI system requirements. Internal Heterogeneity The reference edge AI system architecture is applied to distributed and heterogeneous edge AI systems in the descriptions of various use cases across different industrial domains.
Copyright © 2025 Edge AI Technology Stack 9
Copyright © 2025 Event Organisers 16 The objectives of LoLiPoP IoT (Long Life Power Platforms for Internet of Things) are to develop energy harvesting-based innovative Long Life Power Platforms that enable retrofit of wireless sensor network edge devices for asset tracking, condition and performance monitoring. www.lolipop-iot.eu LoLiPoP IoT EdgeAI-Trust AIMS5.0 SC4EU SC4EU is a unique Chips JU "Innovation Action" project to take the supply chain management of semiconductor production in Europe to a new level. A true demand platform, along with its ontology as a formal description of all information within the chain, facilitates close interaction and smooth, transparent collaboration, making even highly complex supply chains resilient, flexible, and agile. https://sc4.eu/ EdgeAI-Trust addresses an advanced, trustworthy edge AI ecosystem through cuttingedge hardware, software, and tools. The project aims to enhance decentralized EdgeAI operations that are secure, reliable, and sustainable. By integrating AI-based algorithms, devices, and APIs, EdgeAI-Trust fosters interoperability and secure data exchange across diverse platforms. from sensor-actuated devices to cloud systems, all within a dynamic zero trust environment. https://www.edgeai-trust.eu/ AIMS5.0 aims to boost the economy by adopting, extending, and implementing AIenabled HW and SW components and systems across the entire industrial value chain. New technologies from IoT and based on Semantic Web ontologies, ML and AI help European manufacturers to shift from Industry 4.0 to Industry 5.0, creating humancentric workplace conditions and a climate-friendly production. https://aims50.eu/
Copyright © 2025 Supporting Organizations 17 The European Technology Platform on Smart Systems Integration is an industrydriven policy initiative, defining research, development and innovation needs as well as policy requirements related to Smart Systems Integration and integrated Microand Nanosystems. The main objective is to develop a vision and to set up a Strategic Research Agenda. www.smart-systems-integration.org Inside Industry Association is the European Technology Platform for research, design and innovation on Intelligent Digital Systems and their applications. The Association is a membership organisation for the European research and innovation actors with more than 200 members and associates from all over Europe. www.inside-association.eu Chips Joint Undertaking supports research, development, innovation, and future manufacturing capacities in the European semiconductor ecosystem. Launched as part of the Chips for Europe Initiative, it confronts semiconductor shortages and strengthens Europe's digital autonomy, engaging a significant EU, national/regional and private industry funding of nearly €11 billion. https://portal.chips-ju.europa.eu/ EU AENEAS EPoSS INSIDE Chips JU AENEAS standing for Association for European NanoElectronics ActivitieS, is an industrial Association, established in 2006, providing unparalleled networking opportunities, policy influence & supported access to funding to all types RD&I participants in the field of micro and nanoelectronics enabled components and systems. https://aeneas-office.org/