Design and Performance Assessment of GNC Algorithms for Vision Based Rendezvous in Mars Orbit

Paper number

IAC-06-C1.7.06

Author

Mr. Emanuele Di Sotto, DEIMOS Engenharia, Portugal

Coauthor

Dr. Luis F. Penin, DEIMOS Space S.L., Spain

Coauthor

Mr. Tiago S. Hormigo, DEIMOS Engenharia, Portugal

Coauthor

Mr. Juan L. Cano, DEIMOS Space S.L., Spain

Coauthor

Mr. Francisco Câmara, DEIMOS Engenharia, Portugal

Coauthor

Mr. Augusto Caramagno, DEIMOS Space, Spain

Coauthor

Prof. José Manuel Rebordão, Portugal

Coauthor

Mr. Ivo Vieira, Portugal

Coauthor

Mr. Baltazar Parreira, Portugal

Coauthor

Mr. Salvatore Mancuso, European Space Agency/Headquarters, France

Year

2006

Abstract

Autonomous navigation requirement for orbital rendezvous operations is addressed within this paper where Mars Sample Return is considered as reference scenario. The work presented deals with the design and the performance assessment of an autonomous GNC system able to cover the whole proximity range starting from 5 km range up to the attainment of the docking conditions. 

The chaser vehicle is assumed to be equipped with a camera-based sensor in charge of providing measurements of the spacecraft translational and rotational state with respect to the target. These measurements are collected and processed in real time to provide continuous navigation solution on board. It is remarked that this camera-based navigation is performed from a relatively long distance (i.e.5 km), in contrast to current systems (i.e. the ATV-VDM which starts at a relatively close distance of 300 m).

As a reference, the chaser is equipped with a monocular camera while the target spacecraft carries a 3D light pattern on the docking structure, used by the camera navigation system to estimate the relative position and attitude between the two spacecraft. The camera was designed with small focal length and a relative large FOV (10 deg) with good performances at close distance at the expense of moderate range accuracy performance at large distances. The light pattern on the target spacecraft is made of two different patterns: an external and an internal pattern in order to be able to have a pattern always in sight considering the FoV of the Camera. The first pattern was defined in order to minimise the detection error for range, whereas the inner pattern was defined based on minimising the attitude error for the docking phase.

The paper will detail tailored guidance and control algorithms derived for the last phases (far closing, closing and final approach) to meet the mission requirements while fulfilling the constraints posed on the overall mission profile by the camera-based sensor. These requirements and constraints lead to design a specific mission profile based on multi-boost hopping strategies being each boost controlled through a dedicated terminal point guidance scheme. A dedicated coupled control has been designed for the last phase (approach to docking) where translational and rotational dynamics need to be jointly controlled to meet the docking axis acquisition.

The selected navigation scheme is based on the use of a Kalman Filter for estimation of the deviations of the state vector with respect to the reference mission profile. The navigation scheme uses the optical measurements obtained through the camera sensor along with the delta in velocity and in the angular rate provided by the IMU to induce the S/C state vector. The observables are directly fed into the KF. The navigation function is supported by an image processing function that is able to locate and identify the centres of the image spots on the image corresponding to the lights on the light pattern on the target spacecraft. Several modes are defined depending on how many points can be resolved. Results show that pixel/3 accuracy can be guaranteed most of the time for all the light spots in the different range regimes.

The designed GNC system has been integrated in functional engineering simulator (FES) and the performance validated versus the main navigation requirement. Navigation requirements are largely met also at medium/far range distance (1-5 km form the target) where the camera performances are not as good as in the short range. The vision-based GNC system provides reasonable accuracy also around the transition regions where defocusing effects cause the image processing to degrade the spots location precision.

The work presented in this paper was done under the ESA/ESTEC contract for the study of Vision Based Relative Navigation Techniques Framework (VBRNAV). 

Abstract document

IAC-06-C1.7.06.pdf

Manuscript document

IAC-06-C1.7.06.pdf (🔒 authorized access only).

To get the manuscript, please contact IAF Secretariat.