Post-flight Evaluation of Lidar-based Digital Terrain Models for OSIRIS-REx Navigation at Bennu

Abstract

The Origins, Spectral Interpretation, Resource Identification, and Security–Regolith Explorer (OSIRIS-REx) spacecraft spent more than 2 yr characterizing near-Earth asteroid (101955) Bennu. The OSIRIS-REx Laser Altimeter (OLA) was responsible for producing the most accurate reconstruction of the asteroid’s surface—down to a global resolution of around 5 cm with a data precision of ±1.25 cm. However, the best-quality global OLA digital terrain model (DTM), version 21 (v21), was not available for navigation during proximity operations, nor was the utility of this model evaluated for processing images and altimeter data for navigation. The focus of this paper is the post-flight assessment of the final OLA v21 DTM, its performance for navigation-related analysis, and estimates of corrections needed for the DTM and measurement models. We created 15 cm resolution maplets for processing optical navigation (OpNav) data, and 5 cm resolution DTMs for processing altimeter data, to estimate a combined spacecraft trajectory over five phases of the mission. Our estimated corrections to the OLA instrument model produce altimeter data residuals with a precision of 7.12 cm (1σ; one standard deviation from the mean). The OpNav maplets produce image residuals at 0.2 px (1σ) and estimated landmark locations accurate to ±6 cm, outperforming DTM navigation-related performance requirements. Finally, our estimate of the global DTM scale is more precise and within 1.1σ of previously reported values. We find that a slight discrepancy persists between the image and altimeter data, with image data suggesting that the DTM is too small by 0.049%, but nevertheless is exceptional for navigation.

Stereophotoclinometry for OSIRIS-REx Spacecraft Navigation

“…While SPC allows for filtering of landmarks that fall below a certain emission angle threshold, it does not have a filter for illumination angle, so the navigation solutions were potentially susceptible to bias in the orbit normal direction. Leonard et al (2023) further investigate these shape model biases in a postflight assessment of the navigation performance using the final OLA-based DTMs.…”

“…The ALWTG and FDS teams used this new data set in conjunction with data from prior orbital mission phases to identify remaining error sources. FDS performed OD and geophysical parameter-estimation analysis using multiple optical and lidar data types, including classic landmark OpNav, differenced image constraints and altimetric crossovers, and direct altimetry for various combinations of orbital phases, details of which are presented in Leonard et al (2023) and Goossens et al (2021). Landmark OpNav measurements were generated in multiple ways: via the SPC toolkit using both SPC-and OLA-derived maplets, and an alternate independent technique, called surface feature navigation (SFN), which renders surface features in the image frame (Gnam et al 2019;Liounis et al 2019).…”

“…During this investigation, there were differences in the SPC and SFN landmark measurements on the order of 0.2-0.4 px, and it was postulated that the error might be in SPC; however, further research identified a bug in the implementation of SFN, described in Liounis et al (2022). Further FDS investigations have since been conducted, the results of which are described in Goossens et al (2021) and Leonard et al (2023).…”

“…Images were reprocessed after operations were complete using an updated version of the SPC software that included these corrections. Results from this correction are presented in Leonard et al (2023). Additional OSIRIS-REx navigation lessons learned are discussed in Getzandanner et al (2022).…”

OSIRIS-APEX: An OSIRIS-REx Extended Mission to Asteroid Apophis

“…APEX will make DTMs with two complementary techniques: OCAMS-based stereophotoclinometry (SPC) and OLA lidar ranging. Both methods were validated by OREx at Bennu (e.g., Barnouin et al 2019Barnouin et al , 2020Seabrook et al 2019;Daly et al 2020;Al Asad et al 2021;Palmer et al 2022;Seabrook et al 2022;Gaskell et al 2023;Leonard et al 2023). Either technique will produce models that meet the APEX navigation and science objectives, and comparing them will result in a higher-quality DTM of Apophis.…”