Northumbria Uni: First AI Benchmark for Satellite Safety

Northumbria University is leading a new study to develop the first AI benchmark for satellite collision avoidance. This project, called SSA-LaMB, aims to improve how AI systems are used to prevent collisions in space. There are over 40,000 tracked objects in Earth's orbit, leading to more than 144,000 collision-avoidance maneuvers each year. This new benchmark will be the first standardized tool to test how reliably AI systems track space debris and predict collision risks. Currently, no such tool exists. Professor Wai Lok Woo, the lead researcher, states that trustworthy AI in space is an urgent need, as unreliable AI poses real operational risks. The project will make rigorous AI evaluation tools accessible to all researchers. The SSA-LaMB project involves collaborators from Northumbria University, the University of Sheffield, and partners like 3S Northumbria Ltd and ExoAnalytic Solutions. All datasets and tools will be openly released, making space situational awareness research more accessible globally. This initiative matters because it addresses a critical and growing safety concern in Earth's orbit.