On-board cooperative spacecraft relative navigation fusing GNSS with vision

Relative navigation of multiple spacecraft based on the processing Global Navigation Satellite System (GNSS) signals has been successfully performed in Low Earth Orbit (LEO). An accuracy order one millimeter for relative positioning can be achieved post-processing by using Carrier Phase Differential GNSS (CDGNSS) and recovering correct carrier phase integer ambiguity. However, real-time performance without post100761-processing much lower because limited on-board computational resources, signal obstruction, multipath interference induced other orbiting proximity, inaccurate knowledge spacecraft’s attitude. The considerably degraded higher orbits above constellation weaker poor Geometric Dilution Precision (GDOP). On hand, monocular vision is typically not as accurate CDGNSS. In this paper, we review a novel multi-sensor approach that utilizes tight fusion observations measurements which enables fast robust autonomous pose estimation cooperative environment. We consider significant tracking noise high operation characterized signals, visibility, GDOP, low visibility. architecture algorithms are evaluated both simulated Geosynchronous data real GPS collected an urban environment, thereby demonstrating fusing with provide better accuracy, robustness, faster ambiguity resolution than when only, especially conditions.