Terrain Relative Navigation Research Articles

Terrain-Relative Navigation With Neuro-Inspired Elevation Encoding

Terrain-Relative Navigation With Neuro-Inspired Elevation Encoding

InSAR Simulation and Speeded-Up Robust Features Algorithm for Terrain Relative Navigation in PSRs on the Moon

InSAR Simulation and Speeded-Up Robust Features Algorithm for Terrain Relative Navigation in PSRs on the Moon

A High‐Resolution Digital Terrain Model Mosaic of the Mars 2020 Perseverance Rover Landing Site at Jezero Crater

AbstractWe demonstrate the capabilities of a published MADNet monocular height estimation network in producing a refined digital terrain model (DTM) mosaic at 50 cm/pixel resolution for the Mars 2020 Perseverance rover landing site in Jezero crater on Mars. Our approach utilizes the publicly available Mars 2020 Terrain Relative Navigation (TRN) High‐Resolution Imaging Science Experiment (HiRISE) Digital Terrain Model (DTM) mosaic, which was originally created by the United States Geological Survey (USGS) Astrogeology Science Centre. Our resultant HiRISE MADNet DTM mosaic is strictly matched with the original HiRISE TRN DTM and orthoimage mosaics. These mosaics are themselves co‐aligned with the USGS TRN Context Camera (CTX) based DTM and orthoimage mosaics, as well as the ESA/DLR/FUB (European Space Agency/German Aerospace Center/Free University Berlin) High Resolution Stereo Camera (HRSC) level 5 DTM and orthoimage mosaics. In this paper, we provide a brief description of the technical details, and present both visual and quantitative assessments of the refined MADNet HiRISE Jezero DTM mosaic product. This DTM product is now publicly available at http://dx.doi.org/10.17169/refubium-38359.

Moon Diver: Exploring a pit's exposed strata to understand lunar volcanism

Natural pits on the Moon expose deep cross-sections of the lunar maria, enabling direct investigation of the Moon's volcanic history and providing potential access to subsurface lava tubes. The Moon Diver mission concept seeks to explore the Mare Tranquillitatis pit, which exposes the largest wall of bedrock of the mare pits (∼65 m). The concept is enabled by two innovative capabilities: pinpoint landing near the pit and robotic access to its near-vertical wall with an instrument package to examine the elemental chemistry, mineralogy, and morphology of these bedrock layers. Pinpoint landing uses closed-loop guidance with terrain-relative navigation (TRN), which was advanced by Perseverance landing on Mars, to deliver the lander within a 100-m ellipse. The Axel robotic explorer, which remains tethered to the lander, would egress onto the lunar surface and traverse the relatively flat terrain to the pit's funnel entrance. The lander, which is the data link to Earth, also serves as an anchor and provides power and communication to the rover through its tether. The rover is a novel two-wheeled platform with a trailing boom and a spool that pays out the tether as the rover traverses toward the pit. The 300-m long tether is well margined for the rover to scale the pit wall. The rover carries a surface preparation tool and three additional instrument types: (a) three high-resolution cameras for acquiring context images of the near and far walls with the near-wall pair in a stereoscopic configuration, (b) an alpha-particle-X-ray spectrometer (APXS) for elemental composition, and (c) a multi-spectral microscopic imager (MMI) that uses controlled lighting for mineralogy. The surface-preparation tool removes dust and patina from the rock wall by grinding a small area. This tool, the MMI, and the APXS would be deployed from an instrument bay inside the wheel wells. The rover would independently point each instrument at the same target on the wall with millimeter repeatability. Landing shortly after sunrise, the surface mission timeline is just shy of a lunar daytime (14 Earth days). Beyond the primary mission, the rover would be capable of descending from the overhang and peering into the void that may open to a large cave or lava tube. Lunar pits provide an exciting new target for exploration using innovative robotic capabilities that have been tested with integrated science instruments at multiple terrestrial analogue sites including a pit with basaltic layers in Arizona.

Doppler Light Detection and Ranging-Aided Inertial Navigation and Trajectory Recovery

A sensor fusion approach to terrain relative navigation is proposed in this paper. Sensor data from Global Position System (GPS), Inertial Navigation System (INS), and Doppler light detection and ranging (LIDAR) provide a relative state estimate through a novel multiplicative extended Kalman filter. The measurement model of the Doppler LIDAR is derived from first principals and employed by the filter to update the estimates propagated by the GPS/INS data stream. The line-of-sight Doppler velocity measurement provides direct feedback to the velocity states and improves sensitivity to these state estimates. A novel Rauch–Tung–Striebel smoother framework is derived that is compatible with states exhibiting multiplicative error that enables seamless postprocessing of state estimates filtered with a multiplicative extended Kalman filter. This new filter and smoother framework is tested with suborbital flight data from Blue Origin’s New Sheppard ascent and landing missions.

Velocimeter Light-Detection-and-Ranging–Informed Terrain Relative Navigation

Velocimeter light-detection-and-ranging (LIDAR)–informed batch and sequential estimation approaches for terrain relative navigation applications are developed in this paper. State-of-the-art velocimeter LIDAR sensors are capable of delivering simultaneous three-dimensional relative position, relative Doppler velocity, and reflectivity measurements for every point in the field of view. The proposed batch estimation algorithm yields accurate and statistically consistent vehicle velocity estimates in unknown (uncooperative) environments. This work presents novel pointwise relative position and Doppler velocity measurement models that can be utilized in the presence of known (cooperative) environments. Rate estimates obtained by the batch estimator are fused with the pointwise measurement models and classical inertial measurement unit formulations in the form of a multiplicative extended Kalman filter. This Doppler LIDAR-based sensor fusion approach yields informative terrain relative navigation solutions suitable for both cooperative and uncooperative environments. Results from extensive emulation robotics testing performed at NASA Johnson Space Center and Texas A&M’s Land, Air, and Space Robotics laboratory are presented to validate the proposed methodologies.

Implementation of a Map Relative Localization System for Planetary Landing

The Mars 2020 Entry, Descent, and Landing (EDL) system successfully delivered the Perseverance rover to Jezero Crater, the most hazardous landing site ever attempted on Mars. To mitigate the risk of landing hazards, which included cliffs, inescapable dune fields, and rocks, a novel terrain relative navigation system was developed and integrated with the heritage Mars Science Laboratory EDL system. First the hazards were identified using orbital imagery and stored on-board the spacecraft as a hazard map. Then, during parachute descent, the Lander Vision System (LVS) estimated map relative position by fusing landmarks matched between descent imagery and a map of the landing site with inertial measurement unit data. Finally, the position estimate and hazard map were then used by the powered descent system to identify and then fly to the safest and reachable landing target. Postflight analysis indicated that all required systems worked much better than predicted. In particular, the fully autonomous LVS generated a position estimate in 10 s that was in error by only a few meters relative to a 40 m requirement. This paper describes the LVS design, how it was tested before launch, and the LVS performance during EDL.

Optimal Powered Descent Guidance Under Thrust Pointing Constraint

This paper proposes a method for solving a three-dimensional propellant-optimal powered descent guidance problem under a thrust pointing constraint, where the thrust direction is restricted. Incorporating a thrust pointing constraint is as essential as pursuing propellant optimality for modern powered descent guidance, particularly when a lander uses terrain relative navigation sensors. However, the theoretical knowledge of optimal solutions is limited. One of the paper’s contributions is a method for finding the optimal solution under an upward-pointing constraint with an active–inactive–active switching structure. Usually, there is a maximum–minimum–maximum structure for the optimal thrust magnitude. Another contribution is compact propellant-optimal powered descent guidance under a thrust pointing constraint. The guidance may find the solution by establishing analytical models to evaluate switching points from the switching functions and predict the final-state errors. Numerical simulations provide the solutions of lunar pinpoint-landing scenarios, in addition to demonstrating the excellent propellant optimality, smooth convergence, and fast execution.

Initial Positioning of Terrain Relative Navigation Under Pseudo-Peaks Interference

Initial Positioning of Terrain Relative Navigation Under Pseudo-Peaks Interference

Ellipse Crater Recognition for Lost-in-Space Scenario

In the field of deep space exploration, a crater recognition algorithm is key to landing navigation based on craters. When there is only visual information, determining how to identify the crater and provide the initial pose of the lander for a lost-in-space (LIS) scenario is a difficulty in terrain relative navigation (TRN). In this paper, a fast and robust crater recognition method for absolute pose estimation based on projective invariants is proposed, which can provide an accurate initial pose for tracking navigation. First, the method selects navigation craters to establish a small-capacity and high-efficiency crater database, and crater pair serial numbers and projective invariants are stored. Second, our method uses a dynamic threshold to solve the problem that the projective invariants are sensitive to noise. Then, an iterative pyramid algorithm is proposed to quickly filter redundancies. Using a dynamic threshold, the matching rate was increased by at least 10%, and the average processing speed was increased by 40%. When the detection errors of the major and minor axes of the ellipse reached 5%, the detection error of the center point reached 1 pixel, and the tilt angle error reached 5°; the matching rate was still >80%. Finally, the pose was estimated by solving the perspective-n-point (PNP) problem based on the recognized craters. The initial pose error in the simulation environment was less than 2°, and the position error was less than 44 m.

A Novel Visual-Based Terrain Relative Navigation System for Planetary Applications Based on Mask R-CNN and Projective Invariants

In the framework of autonomous spacecraft navigation, this manuscript proposes a novel vision-based terrain relative navigation (TRN) system called FederNet. The developed system exploits a pattern of observed craters to perform an absolute position measurement. The obtained measurements are thus integrated into a navigation filter to estimate the spacecraft state in terms of position and velocity. Recovering crater locations from elevation imagery is not an easy task since sensors can generate images with vastly different appearances and qualities. Hence, several problems have been faced. First, the crater detection problem from elevation images, second, the crater matching problem with known craters, the spacecraft position estimation problem from retrieved matches, and its integration with a navigation filter. The first problem was countered with the robust approach of deep learning. Then, a crater matching algorithm based on geometric descriptors was developed to solve the pattern recognition problem. Finally, a position estimation algorithm was integrated with an Extended Kalman Filter, built with a Keplerian propagator. This key choice highlights the performance achieved by the developed system that could benefit from more accurate propagators. FederNet system has been validated with an experimental analysis on real elevation images. Results showed that FederNet is capable to cruise with a navigation accuracy below 400 meters when a sufficient number of well-distributed craters is available for matching. FederNet capabilities can be further improved with higher resolution data and a data fusion integration with other sensor measurements, such as the lunar GPS, nowadays under investigation by many researchers.

L2AMF-Net: An L2-Normed Attention and Multi-Scale Fusion Network for Lunar Image Patch Matching

The terrain-relative navigation (TRN) method is often used in entry, descent and landing (EDL) systems for position estimation and navigation of spacecraft. In contrast to the crater detection method, the image patch matching method does not depend on the integrity of the database and the saliency of the crater features. However, there are four difficulties associated with lunar images: illumination transformation, perspective transformation, resolution mismatch, and the lack of texture. Deep learning offers possible solutions. In this paper, an L2-normed attention and multi-scale fusion network (L2AMF-Net) was proposed for patch descriptor learning to effectively overcome the above four difficulties and achieve lunar image patch matching accurately and robustly. On the one hand, an L2-Attention unit (LAU) was proposed to generate attention score maps in spatial and channel dimensions and enhance feature extraction. On the other hand, a multi-scale feature self and fusion enhance structure (SFES) was proposed to fuse multi-scale features and enhance the feature representations. L2AMF-Net achieved a 95.57% matching accuracy and excellent performance compared with several other methods in lunar image patch dataset generated in this paper. Experiments verified the illumination, perspective and texture robustness of L2AMF-Net and the validity of the attention module and feature fusion structure.

Square-Root Extended Information Filter for Visual-Inertial Odometry for Planetary Landing

A novel sequential information filter formulation for computationally efficient visual-inertial odometry and mapping is developed in this work and applied to a realistic moon landing scenario. Careful construction of the square-root information matrix, in contrast to the full information or covariance matrix, provides easy and exact mean and covariance recovery throughout operation. Compared to an equivalent extended Kalman filter implementation, which provides identical results, the proposed filter does not require explicit marginalization of past landmark states to maintain constant-time complexity. Whereas measurements to opportunistic visual features only provide relative state information, resulting in drift over time unless a priori mapped landmarks are identified and tracked, the tight coupling of the inertial measurement unit provides some inertial state information. The results are presented in a terrain-relative navigation simulation for both a purely orbital case (with no active propulsion) and a landing case with a constant thrust.

Autonomous Sensor System for Determining Instrument Position Relative to Unknown Surfaces Utilized on Mars Rover

In some applications, a robot arm has to position instrumentation/payloads accurately relative to an unknown surface. This article describes an optical sensor system for doing so. The sensor is based on a number of different image processing techniques combined with optical stimulation. More specifically, the system utilizes a camera, floodlight, structured light illumination, terrain relative navigation (TRN), camera pose estimation, and circle finding. The sensor has been developed for the Planetary Instrument for X-Ray Lithochemistry (PIXL) instrument on NASA’s Mars Perseverance rover. The sensor also has other applications such as exploration of polluted or inaccessible areas or locating a robotic arm relative to an unknown surface. The sensor can aid in placing an instrument relative to the unknown surface with an accuracy of 0.5 mm and maintaining a 3-D position lock utilizing TRN. This article will describe the sensor design and performance.

Absolute Triangulation Algorithms for Space Exploration

Images are an important source of information for spacecraft navigation and for three-dimensional (3-D) reconstruction of observed space objects. Both of these applications take the form of a triangulation problem when the camera has a known attitude and the measurements extracted from the image are line-of-sight (LOS) directions. This work provides a comprehensive review of the history and theoretical foundations of triangulation. A variety of classical triangulation algorithms are reviewed, including a number of suboptimal linear methods (many LOS measurements) and the optimal method of Hartley and Sturm (only two LOS measurements). It is shown that the optimal many-measurement case may be solved without iteration as a linear system using the new linear optimal sine triangulation (LOST) method. Both LOST and the polynomial method of Hartley and Sturm provide the same result in the case of only two measurements. The various triangulation algorithms are assessed with a few numerical examples, including planetary terrain relative navigation, angles-only optical navigation at Uranus, 3-D reconstruction of Notre-Dame de Paris, and angles-only relative navigation.

An Improved ICCP-Based Underwater Terrain Matching Algorithm for Large Initial Position Error

The matching algorithm is a key technology in underwater terrain-relative navigation (TRN). An improved iterative Closest Contour Point (ICCP) matching algorithm is proposed to address the mismatch of the sequence mapping algorithm in the event of the significant initial position error. Based on the rich topographical information provided by the multi-beam echosounder (MBES) in a short time, the conventional ICCP method of searching for target contour points based on single-point matching is improved. In order to reduce the interference of heading error, the terrain distribution feature is modeled as a two-dimensional Gaussian distribution, and a terrain feature descriptor with rotation invariance is established. The similarity between the measured local topographic map and the reference digital topographic map (DTM) is evaluated by the terrain characteristics descriptor to derive the target path. The results of simulation and field tests demonstrate that the improved method can not only realize position matching under large initial errors but also improve the matching accuracy. Furthermore, the simulation and field tests show that the improved ICCP algorithm works well in both flat and rough areas.

Synthetic-Aperture-Radar–Based Spacecraft Terrain Relative Navigation

A spacecraft terrain relative navigation system for low elevation orbit satellites is presented relying on measurements obtained from a synthetic aperture radar (SAR) that are fused with inertial measurements and attitude determination system data in a multiplicative extended Kalman filter. Techniques such as autofocusing and geocoding are employed to retrieve information from SAR images for incorporation into the navigation filter. Monte Carlo simulation results are presented in which actual filter performance is compared to predicted filter performance. Specifically, two test cases, with varied initial SAR antenna misalignment uncertainty, show that the SAR-based terrain relative navigation system is a valid approach to GPS-denied navigation.

High-Resolution Feature Pyramid Network for Automatic Crater Detection on Mars

Crater detection is widely used in terrain relative navigation, which can help scientists target a spacecraft’s position and estimate the age of a planet. However, high-performance crater detection still remains a very challenging problem because of the complex data distribution, especially when there are lots of small craters. To address these issues, we present an end-to-end deep convolution neural network called high-resolution feature pyramid network (HRFPNet) to well detect impact craters. In the network, a new adaptive anchor calculation and label assignment (AACLA) algorithm is designed to solve the problem of “anchor-sensitive” small craters that cannot be balanced sampled for training, which affects the detection results of small craters. Then, a new backbone with feature aggregation module is presented to enhance the feature extraction capabilities and to preserve the features of small craters in deep neural network, and a new regression loss function called balanced regression loss (BRL) is also applied for the coordinate regression of small craters, which avoid the inaccurate size prediction of small objects due to the large difference of the regression loss value between the small craters and other craters. In addition, we build a crater detection dataset called the Mars day crater detection (MDCD) dataset that contains 500 images with 12 000 craters, which can be downloaded from the website: <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">http://doi.org/10.5281/zenodo.4750929/</uri> . We conduct extensive experiments based on the public and MDCD datasets, and the results show that the proposed network achieves the state-of-the-art results; especially, it achieves high performance in detecting small craters.

Lunar Crater Identification in Digital Images

It is often necessary to identify a pattern of observed craters in a single image of the lunar surface and without any prior knowledge of the camera’s location. This so-called “lost-in-space” crater identification problem is common in both crater-based terrain relative navigation (TRN) and in automatic registration of scientific imagery. Past work on crater identification has largely been based on heuristic schemes, with poor performance outside of a narrowly defined operating regime (e.g., nadir pointing images, small search areas). This work provides the first mathematically rigorous treatment of the general crater identification problem. It is shown when it is (and when it is not) possible to recognize a pattern of elliptical crater rims in an image formed by perspective projection. For the cases when it is possible to recognize a pattern, descriptors are developed using invariant theory that provably capture all of the viewpoint invariant information. These descriptors may be pre-computed for known crater patterns and placed in a searchable index for fast recognition. New techniques are also developed for computing pose from crater rim observations and for evaluating crater rim correspondences. These techniques are demonstrated on both synthetic and real images.

Making an Onboard Reference Map From MRO/CTX Imagery for Mars 2020 Lander Vision System

AbstractThe Mars 2020 rover, Perseverance, landed in Jezero crater (18.4663°N, 77.4298°E) on February 18, 2021 to collect samples from Mars that could be returned to Earth by a future Mars Sample Return campaign. While providing a rich sampling opportunity, Jezero also contains numerous landing hazards including scarps, canyons, mesas, dune fields, rock fields, and smaller craters. The Mars 2020 onboard inertial navigation system, which is the same as that used by the Mars Science Laboratory (MSL), only provides a very coarse inertially propagated position, which can have error as large as 3.2 km. The Lander Vision System (LVS) was added to Mars 2020 to reduce this position knowledge error to less than 40 m with respect to an on‐board reference map of the landing area. LVS uses a reference map on board to compare with the descent images for lander localization during the terminal stage of Entry, Descent and Landing (EDL). Because the reference map is used directly during EDL, it is critical that it have as little spatial and photometric error as possible. Photometrically, it should resemble as much as possible the real descent images to allow reliable terrain matching. Spatially, it should match as faithfully as possible the real, underlying terrain and contribute minimal error to the final localization solution. On February 18, 2021, the LVS reference map passed its ultimate test. The LVS system executed Terrain Relative Navigation (TRN) flawlessly based on the LVS reference map. In addition to its use for TRN, the local hazard map is also registered to the LVS reference map so that the safe target selection (STS) system can select a safe and reachable landing site. The final error between the site targeted by the STS system and the real landing site is estimated at only 5 m. The exceptional performance of LVS indicates that the reference map met mission requirements with comfortable margin. LVS on Mars 2020 represents the first ever use of a reference map during spacecraft EDL. This breakthrough will have profound implications for future lander missions and the scope of scientific inquiry they are able to address. In this paper, we will describe the necessary precursor steps to building the Jezero Crater LVS map, including the methodology to dejitter Context Imager (CTX) images, improve the CTX sensor model, as well as the process used to validate the LVS map accuracy.