Skip to content Skip to navigation
Vision-based Monocular Vision Navigation for Close-Range Spacecraft Rendezvous
                                

                                    PRISMA space imagery (credits: OHB/DTU)

Monocular vision navigation is considered an enabling technology for present and future formation-flying and on-orbit-servicing missions. Indeed simple instruments such as star trackers or infrared cameras can be modified for increased dynamic range to accurately navigate with respect to a target space vehicle at low cost from virtually zero to several tens of kilometers.
This research addresses the implementation of novel vision systems for estimating the relative position and relative orientation (i.e., pose) of a space resident object using two-dimensional images collected on-board with none or little a-priori knowledge. Although many authors have addressed this problem through computer vision methods for terrestrial applications, several problems are encountered when dealing with space navigation and are investigated in this work. Typically the initialization procedure of a pose estimator is not able to handle lost-in-space configurations without a-priori knowledge. In addition space imagery is characterized by low carrier-to-noise ratio and high image contrasts, which cause false or partial detections of physical edges and the extraction of a restricted amount of feature points. Furthermore most artificial satellites have symmetric polyhedral shapes. Although these are easy to model, the perspective equations provide several ambiguous pose solutions in the presence of high symmetry which cannot be neglected in the vision system. Finally the unknown correspondences between image and model features result in a large search space for ambiguity resolution, and thus in an unacceptable computational load.
This work seeks the development of cutting-edge techniques for spaceborne optical navigation and their rigorous validation through actual space imagery and high-fidelity hardware-in-the-loop optical stimulators. So far this research has resulted in one of the few on-orbit demonstration of noncooperative rendezvous published in literature (in the frame of the PRISMA mission). Current efforts are put into the generalization and improvement of all functional elements of a pose estimator, from image collection and processing to edge and feature points detection, from initialization to spacecraft modeling, from model matching to pose estimation and refinement. Probabilistic approaches based on perceptual organization and random sample consensus are sought after. Analytical and numerical solutions of the true perspective equations are investigated.
 
Recent Publications (2013-2014)
 
D’Amico S., Benn M., and Jørgensen J.L.;
Pose Estimation of an Uncooperative Spacecraft from Actual Space Imagery;
International Journal of Space Science and Engineering (2013).
Also presented at
5th International Conference on Spacecraft Formation Flying Missions and Technologies (SFFMT), 29-31 May 2013, Munich, Germany (2013).

Gaias G., D'Amico S., and Ardaens J.-S.;
Angles-only Navigation to a Non-Cooperative Satellite using Relative Orbital Elements;
AIAA Journal of Guidance, Control, and Dynamics (2013).
Also presented at
AIAA/AAS Astrodynamics Specialist Conference, 13-16 Aug. 2012, Minneapolis, USA (2012).

D’Amico S., Ardaens J.-S., Gaias G., Benninghoff H., Schlepp B., Jørgensen J.L.;
Noncooperative Rendezvous using Angles-only Optical Navigation: System Design and Flight Results;
AIAA Journal of Guidance, Control, and Dynamics, 36(6) 1576-1595, doi: 10.2514/1.59236 (2013).

GPS-based Absolute and Relative Navigation for Distributed Space Systems
                                

                      Artist illustration of carrier-phase differential GPS (credits: DLR)

A fundamental need of advanced distributed space systems is the determination of the relative motion between individual satellites either post-facto on-ground, or in near real-time on-board. The former task is typically required to support mission operations, mission planning, for scientific data processing, and for verification purposes, whereas the latter one is required for autonomous operations at bus and payload levels. Especially when the multi-satellite system is deployed in low Earth orbit, differential Global Navigation Satellite System (GNSS) represents the primary technique for determining the relative position and velocity of the participating spacecraft.
Similar to terrestrial applications, spaceborne relative navigation benefits from a high level of common error cancellation. Furthermore the integer nature of double-difference carrier-phase ambiguities can be exploited through carrier phase differential GNSS (CDGNSS) techniques. Both aspects enable a substantially higher relative accuracy than can be achieved in single-spacecraft navigation. This research aims at pushing forward the frontiers of GNSS absolute and relative navigation for spaceborne applications such as cooperative spacecraft formation flying and on-orbit servicing. Investigations include the assessment of spaceborne GNSS receivers and antennas, the refinement of dynamics and measurement models for absolute and relative navigation using single- or multiple-frequency measurements, the analysis of linear and non-linear estimation schemes for real-time and offline applications.
This work has produced one of the most accurate and versatile operational spaceborne CDGNSS subsystems published in literature, namely the PRISMA GNSS navigation system. It represents the primary formation-flying sensor on-board, and its key objective is the continuous provision of accurate and reliable relative motion information at all times during the mission. After more than two years in orbit, the demonstrated overall performance is well below 10 cm and 1 mm/s (3D, R.M.S.) for relative position and velocity respectively.
Despite the good results several aspects need to be further studied to enable the ultimate accuracy, robustness, and reliability required by future autonomous multi-satellite systems. These include integer ambiguity resolution in real-time on-board to achieve millimeter level positioning, modeling and rejection of multi-path effects at close range (e.g., for cooperative rendezvous&docking, space structure assembly), near-omni-directional visibility for safe mode or inspection operations. Future research will address GNSS navigation for nanosatellites and for high-altitudes missions. These scenarios offer new challenges due to the scarcity of on-board resources on the one hand, and due to the weakness, and poor geometry of the radio-frequency signals on the other hand.
 
Recent Publications (2013-2014)
 
Ardaens J.-S., D’Amico S., Sommer J.;
GPS Navigation System for Challenging Close-Proximity Formation-Flight;
24th International Symposium on Spaceflight Dynamics, 5-9 May 2014, Laurel, USA (2014).

Ardaens, J.-S., D’Amico S., Cropp A.;
GPS-Based Relative Navigation for the Proba-3 Formation Flying Mission;
Acta Astronautica, 91, 341-355. DOI 10.1016/j.actaastro.2013.06.025 (2013).

Montenbruck O., D'Amico S.;
GPS Based Relative Navigation;
Chap 5, pp. 185-223. In: D' Errico M. (Ed.)
Distributed Space Missions for Earth System Monitoring, Space Technology Library, 2013, Volume 31, Part 2, 185-223.
doi 10.1007/978-1-4614-4541-8_5 (2013)
 
Angles-Only Navigation for Space Rendezvous using Relative Orbital Elements
                                

                       Angles-Only Navigation for the mSTAR mission (credits: SLAB)

The capability to navigate with respect to other resident space objects using simple, low-cost sensors such as an optical camera enables a variety of strategic distributed space missions ranging from non-cooperative on-orbit servicing to distributed science applications. Accordingly, this research focuses on improving the capability to autonomously approach, identify, and/or rendezvous with a passive target from far-range using safe and fuel efficient algorithms which rely only on camera-based measurements.
The angles-only navigation research at SLAB builds from the strong foundation developed in the ARGON experiment phase of the PRISMA mission by leveraging the advanced Relative Orbital Element (ROE) state to enable seamless inclusion of perturbations, greater insight into the relative orbit geometry, robust collision-avoidance, and intuitive maneuver-planning. However, the angles-only navigation approach used in ARGON relied on known orbit maneuvers to remedy the well-documented observability issues arising from a lack of direct range information. Instead, current research efforts at SLAB have been focused on removing the reliance upon maneuvers for observability improvement to simplify the overall navigation procedure. In recently published work by SLAB researchers, a comprehensive observability assessment revealed that including the secular J2 perturbation effects in the dynamics model as well as nonlinearities in the measurement model vastly improves the system observability. Subsequently, it was shown for the first time that a maneuver-free angles-only navigation architecture subject to realistic sensor constraints and simulation conditions is able to produce full-state estimation at the same level of accuracy as previously published approaches which relied on maneuvers.
Current efforts are seeking to improve upon this work by extending the application to orbits of arbitrary eccentricity and by considering additional filter designs. In the long-term, the angles-only navigation architectures developed through this research will be used to meet the requirements posed by future distributed space systems under development at Stanford University. Namely, the proposed mDOT mission will rely heavily on angles-only techniques for precision relative navigation of a distributed occulter and telescope spacecraft formation, and the mSTAR mission launching in 2020 will demonstrate fully-autonomous vision-based navigation and rendezvous of a microsatellite with an ejected non-cooperative nanosatellite in low Earth orbit.
 
Recent Publications (2013-2016)
 
Sullivan, J., Koenig, A., and D’Amico, S.;
Improved Maneuver-Free Approach to Angles-Only Navigation for Space Rendezvous;
AAS/AIAA Spaceflight Mechanics Conference (2016).

Gaias G., D'Amico S., and Ardaens J.-S.;
Angles-only Navigation to a Non-Cooperative Satellite using Relative Orbital Elements;
AIAA Journal of Guidance, Control, and Dynamics (2013).
Also presented at
AIAA/AAS Astrodynamics Specialist Conference, 13-16 Aug. 2012, Minneapolis, USA (2012).

D’Amico S., Ardaens J.-S., Gaias G., Benninghoff H., Schlepp B., Jørgensen J.L.;
Noncooperative Rendezvous using Angles-only Optical Navigation: System Design and Flight Results;
AIAA Journal of Guidance, Control, and Dynamics, 36(6) 1576-1595, doi: 10.2514/1.59236 (2013).