Satellites capable of moving, understanding their surroundings, and making decisions completely on their own — this is the goal of project A3, coordinated by the HEAPLab of the Department of Electronics, Information and Bioengineering – Politecnico di Milano, in collaboration with the Department of Aerospace Science and Technology and the aerospace company Leonardo.

The project aims to develop an intelligent guidance, navigation, and control system for space platforms capable of performing orbital servicing and assembly operations autonomously, without human intervention, through the use of hybrid artificial intelligence. This type of AI enables the satellite to reason on multiple levels, adapting in real time to complex and rapidly changing scenarios.

One of the main challenges is to ensure that satellites can quickly reconfigure their trajectory, move precisely near other objects, and safely manage delicate operations, such as assembling large structures in space — even in the presence of unpredictable variables like shape, mass, and dynamics. Project A3 will start with simpler scenarios, where a single satellite approaches a “target” object, and will evolve toward more complex situations involving multiple satellites working together in formation. These systems will be supported by advanced environmental interpretation technologies, already used on Earth, now adapted to space to significantly increase onboard autonomy.

The team plans an extensive campaign of laboratory and simulated environment testing, aiming to reach a Technology Readiness Level (TRL) between 3 and 4. Among the innovations is a test platform based on RISC-V architecture, enhanced with FPGA-based accelerators — a tangible demonstration that commercial off-the-shelf components can also deliver high performance in space applications.

Project A3 represents a major step toward a future where satellites can collaborate autonomously, helping reduce the risks, time, and costs associated with space missions.