Project description
Demonstrating the potential of edge-AI technology on a satellite
Satellites are crucial for Earth observation, attracting significant investment from both the private and public sector and potentially reshaping global economics. However, current data management infrastructure cannot sufficiently maximise the value of the increasing amount of data captured. To address this, the EU-funded Edge SpAIce project plans to develop an efficient approach to deploying deep neural networks (DNNs) at the edge for more effective data management, targeting continuous data flow between capture and processing. Despite challenges such as high computational power requirements and complex DNN architectures, Edge SpAIce will optimise AI execution, enabling its compatibility with various on-board satellite hardware. The ultimate goal is to demonstrate the potential of edge-AI technology by deploying a DNN for remote monitoring of marine plastic litter on a satellite.
Objective
"Satellites have become one of most prominent technologies for Earth Observation, progressively attracting large investments from both the private and public sector and potentially becoming the next life-changing trend in world economics. Resulting from technological advancements and gradual reduction in manufacturing and in-orbit deployment costs, data captured drastically increased along, revealing significant impairments in current data management infrastructure to maximise associated added value. With the purpose of achieving a more efficient data management to further approach novel EO applications requiring continuous data flow between capturing and processing, Deep Neural Networks (DNNs) deployment ""at the edge"" has been investigated as valid approach to allow autonomous and reliable data payload and latency reduction while keeping high added value from data captured. However, deployment of accurate DNNs models present several limitations, with the major ones being high computational power and cumbersome architectures. In this project proposal, Edge SpAIce develops an extremely efficient approach to resize complex DNNs while ensuring compatibility requirements for on-board satellites hardware are met. With this scope, Edge SpAIce will further target a challenging demonstration of Edge-AI potential by design and deployment of a DNN for marine plastic litter remote monitoring from in-orbit representative satellite, paving the way towards next generation EO and moving European leadership in the global space market."
Fields of science
- natural sciencesphysical sciencesastronomyobservational astronomyoptical astronomy
- social scienceseconomics and businesseconomics
- engineering and technologymechanical engineeringvehicle engineeringaerospace engineeringsatellite technology
- natural sciencescomputer and information sciencesartificial intelligencecomputational intelligence
Keywords
Programme(s)
Topic(s)
Funding Scheme
HORIZON-IA - HORIZON Innovation ActionsCoordinator
78100 SAINT-GERMAIN-EN-LAYE
France
The organization defined itself as SME (small and medium-sized enterprise) at the time the Grant Agreement was signed.