Preliminary Design and Testing of Multi-sensor Relative Navigation Architectures for the SpEye CubeSat Mission

Achieving robust autonomous navigation capabilities is crucial for On-Orbit Servicing missions aiming to extend the lifespan and eventually enhance the functionalities of satellites, thereby optimizing the use of orbital assets and ensuring a sustainable future use of outer space. In this context, the Italian Space Agency (ASI) has funded the Space Eye (SpEye) mission, which is part of the “Alcor” program and is currently in Phase B. SpEye is a two-satellite demonstration mission conceived to develop and validate in orbit key technologies and methodologies to enable formation flying, close-proximity inspection, and rendezvous. The space segment includes a CubeSat-based free-flying nanosatellite that will perform a set of on-orbit demonstrations with its deployer, i.e., the ION CubeSat carrier developed by D-Orbit, either acting as a cooperative spacecraft thanks to an active communication link to exchange GNSS raw data or as an uncooperative target. This paper focuses on the relative navigation sub-system. In particular, it discusses the preliminary design, development, and testing of the relative navigation function for two operative modes: an Electro-Optical-only mode featuring the exclusive use of Electro-Optical sensors, namely a multispectral camera and a laser range finder to provide uncooperative rendezvous capabilities; a multi-sensor mode featuring the combined use of visual and laser data with Carrier Phase Differential Global Navigation Satellite System (CPDGNSS) measurements, thus allowing relative navigation also in eclipse conditions. Both modes exploit a sensor fusion scheme to integrate visual-based observables, laser ranging data, and CDGNSS measurements in a loosely coupled fashion into a Multiplicative Extended Kalman Filter. The visual processing pipeline is conceived to deal with a wide range of target distances by adaptively extracting from the acquired images either the complete pose parameters, thus enabling roto-translational relative state estimation, or only its line-of-sight that, combined with the laser ranging measurements, can be used for translation-only relative state estimation. The problem of autonomously switching between these two operative sub-modes is addressed by checking a thresholding criterion on the apparent size of the target measured on the image plane over a sequence of successive frames. The overall architecture incorporates measurements preprocessing and consistency check steps to enhance robustness. Performance assessment is carried out within a numerical environment realistically reproducing the satellites’ orbital and rotational dynamics as well as the output of absolute and relative navigation sensors, including a synthetic image generator to adequately test the capability to extract from images the desired visual observables.