Onboard Implementation and Validation of RTK-Based Relative Navigation System for CubeSats

This paper presents the first onboard Real-Time Kinematic (RTK)-based relative navigation system designed specifically for CubeSat rendezvous and docking missions. Traditionally applied in larger satellites, GPS-based relative navigation has been optimized for low-Earth orbit (LEO), where minimal signal interference allows precise navigation without additional sensors. However, previous systems faced limitations in real-time resolution of integer ambiguities inherent in GPS carrier-phase measurements, particularly in response to the rapidly changing satellite constellations in LEO. To overcome these issues, this paper adapts a conventional single-station RTK system for use in LEO. An efficient real-time Recursive Ambiguity Filter (RAF) is proposed to reduce the ambiguity search space, replacing the extended Kalman filter. Optimized for CubeSats with constrained computational resources, the algorithm was validated through ground-based tests using commercial GPS receivers and patch antennas. The proposed system performs real-time integer ambiguity resolution, achieving cm-level navigation accuracy and ensuring a 100% success-fix rate. Furthermore, it resolves integer ambiguities within 1 s using the CubeSat's on-board computer (OBC), marking a significant advancement in CubeSat navigation technology by enabling precise satellite positioning for rendezvous missions without requiring additional expensive hardware.