Advancing Energy-Efficient Embedded AI for a Smarter, Sustainable Europe
Duration: 06/2023 – 06/2026 – 36 months
Funding: Horizon Europe
Partners: COMMISSARIAT À L’ÉNERGIE ATOMIQUE ET AUX ENERGIES ALTERNATIVES (CEA), NANOXPLORE, THALES, KOMPAI ROBOTICS, STICHTING IMEC NEDERLAND, GRAI MATTER LABS BV, TECHNISCHE UNIVERSITEIT DELFT (TU Delft),ALMENDE BV, SENSITY B.V., SORAMA BV, SILICON AUSTRIA LABS GMBH, SPIKI GmbH, TTTECH AUTO AG, TTTECH COMPUTERTECHNIK AG, FH-CAMPUS WIEN – VEREIN ZUR FÖRDERUNG DES FACHHOCHSCHUL-STUDIUMS,STMICROELECTRONICS SRL, DEEPSENSING S.R.L., HIGH TECHNOLOGY SYSTEMS s.r.l.,FONDAZIONE BRUNO KESSLER,UNIVERSITÀ DEGLI STUDI DI CATANIA, UNIVERSITÀ DEGLI STUDI DI PALERMO,UNIVERSITÀ DEGLI STUDI DI MESSINA,FRAUNHOFER GESELLSCHAFT ZUR FÖRDERUNG DER ANGEWANDTEN FORSCHUNG E.V., INFINEON TECHNOLOGIES AG, CODASIP GMBH, BRANDENBURGISCHE TECHNISCHE UNIVERSITÄT COTTBUS (BTU)
Neurokit2e
General project description
Europe faces a critical challenge in the domain of embedded Artificial Intelligence (AI): a dependency on American and Chinese tools and technologies, which limits its competitive edge in the rapidly growing embedded AI market.
This is compounded by the fragmented ecosystem of AI hardware and software, hindering easy design, optimization, and implementation of AI applications on constrained hardware.
The NEUROKIT2E project offers a transformative solution through the development of an open-source, deep learning platform for embedded hardware tailored to European industries. Powered by the N2D2 platform developed by the CEA, the project will:
- Enable seamless design, optimization, and deployment of Neural Networks on energy-efficient embedded systems.
- Promote integration with European hardware technologies, enhancing performance, reliability, and cost-efficiency.
- Support scalability to address diverse AI applications in edge computing and IoT.
- Reduce dependency on external technologies and advance Europe’s position in the competitive landscape of embedded AI.
By fostering a robust European value chain, NEUROKIT2E will reduce dependency on foreign technologies, ensure sovereignty in AI innovation, and position Europe as a leader in the embedded AI sector.
Contribution
Almende will play a key role in the NEUROKIT2E project through its Crownstone division.
Almende’s team is developing a neural network infrastructure integrated with the project’s open-source components, enhancing energy-efficient embedded AI systems.
This neural network is capable of performing device identification based on power usage data from various household appliances.
Leveraging the open-source framework Aidge, Almende is optimizing the neural network for deployment on embedded systems, ensuring efficient and scalable AI solutions.
Additionally, Almende is actively collaborating with project partners to integrate and refine the framework components of the NEUROKIT2E platform, making it accessible to diverse industries.
Beyond technical development, Almende plays a crucial role in a key Use Case of the project where in partnership with two Dutch organizations, it is contributing to cutting-edge applications, including fall prevention through gait analysis, environmental sound filtering, and real-time indoor user localization.
This work enhances safety and quality of life for users, showcasing Almende’s expertise in AI-driven IoT solutions.
Results
The neural network developed by Almende for device identification will significantly enhance the functionality of Crownstone’s products, enabling more precise and intelligent power usage monitoring.
In addition, Almende is introducing cutting-edge mechanisms for far-edge adaptivity, enhancing neural network processing directly within far-edge Crownstone nodes.
By leveraging its micro-app architecture, Almende ensures that these nodes operate efficiently while being managed by an in-edge hub device. This approach enables real-time, low-latency decision-making, and improved scalability for smart home and industrial IoT applications.
These innovations will lead to smarter automation, enhanced user experiences, and a more adaptive IoT ecosystem.
More info?
Need specific information regarding the project? Please contact our senior consultant for more information.
