An Autonomous, End-to-End, Convex-Based Framework for Close-Range Rendezvous Trajectory Design and Guidance with Hardware Testbed Validation

Researchers have developed CORTEX, an autonomous AI framework for satellite close-range rendezvous and docking. It combines deep-learning perception with convex optimization for real-time trajectory guidance, including abort logic for failures. In high-fidelity simulations with thrust errors, it achieved terminal docking errors of 36.85 ± 44.46 mm. Crucially, hardware tests on a planar air-bearing testbed demonstrated even higher precision, with terminal errors of just 8.09 ± 5.29 mm in position.