arXiv subjects

Simone D'Amico

Publications and source records attributed to Simone D'Amico.

50 records · Page 3Linked to original sources

arXiv2021

SPEED+: Next-Generation Dataset for Spacecraft Pose Estimation across Domain Gap

Autonomous vision-based spaceborne navigation is an enabling technology for future on-orbit servicing and space logistics missions. While computer vision in general has benefited from Machine Learning (ML), training and validating spaceborne ML models are extremely challenging due to the impracticality of acquiring a large-scale labeled dataset of images of the intended target in the space environment. Existing datasets, such as Spacecraft PosE Estimation Dataset (SPEED), have so far mostly relied on synthetic images for both training and validation, which are easy to mass-produce but fail to resemble the visual features and illumination variability inherent to the target spaceborne images. In order to bridge the gap between the current practices and the intended applications in future space missions, this paper introduces SPEED+: the next generation spacecraft pose estimation dataset with specific emphasis on domain gap. In addition to 60,000 synthetic images for training, SPEED+ includes 9,531 hardware-in-the-loop images of a spacecraft mockup model captured from the Testbed for Rendezvous and Optical Navigation (TRON) facility. TRON is a first-of-a-kind robotic testbed capable of capturing an arbitrary number of target images with accurate and maximally diverse pose labels and high-fidelity spaceborne illumination conditions. SPEED+ is used in the second international Satellite Pose Estimation Challenge co-hosted by SLAB and the Advanced Concepts Team of the European Space Agency to evaluate and compare the robustness of spaceborne ML models trained on synthetic images.

cs.CV↗

arXiv2021

Analytical Process Noise Covariance Modeling for Absolute and Relative Orbits

This paper develops new analytical process noise covariance models for both absolute and relative spacecraft states. Process noise is always present when propagating a spacecraft state due to dynamics modeling deficiencies. Accurately modeling this noise is essential for sequential orbit determination and improves satellite conjunction analysis. A common approach called state noise compensation models process noise as zero-mean Gaussian white noise accelerations. The resulting process noise covariance can be evaluated numerically, which is computationally intensive, or through a widely used analytical model that is restricted to an absolute Cartesian state and small propagation intervals. Moreover, mathematically rigorous, analytical process noise covariance models for relative spacecraft states are not currently available. To address these limitations of the state of the art, new analytical process noise covariance models are developed for state noise compensation for both Cartesian and orbital element absolute state representations by leveraging spacecraft relative dynamics models. Two frameworks are then presented for modeling the process noise covariance of relative spacecraft states by assuming either small or large interspacecraft separation. The presented techniques are validated through numerical simulations.

math.DS↗

arXiv2021

Adaptive and Dynamically Constrained Process Noise Estimation for Orbit Determination

This paper introduces two new algorithms to accurately estimate the process noise covariance of a discrete-time Kalman filter online for robust orbit determination in the presence of dynamics model uncertainties. Common orbit determination process noise techniques, such as state noise compensation and dynamic model compensation, require offline tuning and a priori knowledge of the dynamical environment. Alternatively, the process noise covariance can be estimated through adaptive filtering. However, many adaptive filtering techniques are not applicable to onboard orbit determination due to computational cost or the assumption of a linear time-invariant system. Furthermore, existing adaptive filtering techniques do not constrain the process noise covariance according to the underlying continuous-time dynamical model, and there has been limited work on adaptive filtering with colored process noise. To overcome these limitations, a novel approach is developed which optimally fuses state noise compensation and dynamic model compensation with covariance matching adaptive filtering. This yields two adaptive and dynamically constrained process noise covariance estimation techniques. Unlike many adaptive filtering approaches, the new techniques accurately extrapolate over measurement outages and do not rely on ad hoc methods to ensure the process noise covariance is positive semi-definite. The benefits of the proposed algorithms are demonstrated through two case studies: an illustrative linear system and the autonomous navigation of two spacecraft orbiting an asteroid.

eess.SP↗

arXiv2021

Autonomous Angles-Only Multi-Target Tracking for Spacecraft Swarms

This paper presents a new algorithm for autonomous multitarget tracking of resident space objects using optical angles-only measurements from a spaceborne observer. To enable autonomous angles-only navigation of spacecraft swarms, an observer must be able to identify and track multiple known or unknown target space objects in view, without reliance on a-priori relative orbit knowledge. Extremely high assignment precision is necessary with low measurement frequencies and limited computational resources. The new "Spacecraft Angles-only MUltitarget tracking System" (SAMUS) algorithm has been developed to meet these objectives and constraints. It combines domain-specific modeling of target kinematics with multi-hypothesis techniques to autonomously track multiple unknown targets using only sequential camera images. A measurement transform ensures that target motion in the observer reference frame follows consistent parametric models; curve fitting is used to predict track behavior; and kinematically-derived track gating and scoring criteria are applied to improve the efficiency and accuracy of the multi-hypothesis approach. Monte-Carlo testing with high-fidelity simulations demonstrates close to 100% measurement assignment precision and strong recall across a range of multi-spacecraft formations, in both near-circular and eccentric orbits. Tracking is maintained in the presence of eclipse periods, significant measurement noise, and partially known swarm maneuvers. A comparison to other tracking algorithms reveals strong advantages in precision, robustness and computation time, crucial for spaceborne angles-only navigation.

eess.SP↗

arXiv2021

Starshade Rendezvous: Exoplanet Sensitivity and Observing Strategy

Launching a starshade to rendezvous with the Nancy Grace Roman Space Telescope would provide the first opportunity to directly image the habitable zones of nearby sunlike stars in the coming decade. A report on the science and feasibility of such a mission was recently submitted to NASA as a probe study concept. The driving objective of the concept is to determine whether Earth-like exoplanets exist in the habitable zones of the nearest sunlike stars and have biosignature gases in their atmospheres. With the sensitivity provided by this telescope, it is possible to measure the brightness of zodiacal dust disks around the nearest sunlike stars and establish how their population compares to our own. In addition, known gas-giant exoplanets can be targeted to measure their atmospheric metallicity and thereby determine if the correlation with planet mass follows the trend observed in the Solar System and hinted at by exoplanet transit spectroscopy data. In this paper we provide the details of the calculations used to estimate the sensitivity of Roman with a starshade and describe the publicly available Python-based source code used to make these calculations. Given the fixed capability of Roman and the constrained observing windows inherent for the starshade, we calculate the sensitivity of the combined observatory to detect these three types of targets and we present an overall observing strategy that enables us to achieve these objectives.

astro-ph.IM↗

arXiv2021

Starshade Rendezvous: Exoplanet Orbit Constraints from Multi-Epoch Direct Imaging

The addition of an external starshade to the {\it Nancy Grace Roman Space Telescope} will enable the direct imaging of Earth-radius planets orbiting at $\sim$1 AU. Classification of any detected planets as Earth-like requires both spectroscopy to characterize their atmospheres and multi-epoch imaging to trace their orbits. We consider here the ability of the Starshade Rendezvous Probe to constrain the orbits of directly imaged Earth-like planets. The target list for this proposed mission consists of the 16 nearby stars best suited for direct imaging. The field of regard for a starshade mission is constrained by solar exclusion angles, resulting in four observing windows during a two-year mission. We find that for habitable-zone planetary orbits that are detected at least three times during the four viewing opportunities, their semi-major axes are measured with a median precision of 7 mas, or a median fractional precision of 3\%. Habitable-zone planets can be correctly identified as such 96.7\% of the time, with a false positive rate of 2.8\%. If a more conservative criteria is used for habitable-zone classification (95\% probability), the false positive rate drops close to zero, but with only 81\% of the truly Earth-like planets correctly classified as residing in the habitable zone.

astro-ph.IM↗

arXiv2020

Fast Algorithm for Fuel-Optimal Impulsive Control of Linear Systems with Time-Varying Cost

This paper presents a new fast and robust algorithm that provides fuel-optimal impulsive control input sequences that drive a linear time-variant system to a desired state at a specified time. This algorithm is applicable to a broad class of problems where the cost is expressed as a time-varying norm-like function of the control input, enabling inclusion of complex operational constraints in the control planning problem. First, it is shown that the reachable sets for this problem have identical properties to those in prior works using constant cost functions, enabling use of existing algorithms in conjunction with newly derived contact and support functions. By reformulating the optimal control problem as a semi-infinite convex program, it is also demonstrated that the time-invariant component of the commonly studied primer vector is an outward normal vector to the reachable set at the target state. Using this formulation, a fast and robust algorithm that provides globally optimal impulsive control input sequences is proposed. The algorithm iteratively refines estimates of an outward normal vector to the reachable set at the target state and a minimal set of control input times until the optimality criteria are satisfied to within a user-specified tolerance. Next, optimal control inputs are computed by solving a quadratic program. The algorithm is validated through simulations of challenging example problems based on the recently proposed Miniaturized Distributed Occulter/Telescope small satellite mission, which demonstrate that the proposed algorithm converges several times faster than comparable algorithms in literature.

math.OC↗

arXiv2020

Satellite Pose Estimation Challenge: Dataset, Competition Design and Results

Reliable pose estimation of uncooperative satellites is a key technology for enabling future on-orbit servicing and debris removal missions. The Kelvins Satellite Pose Estimation Challenge aims at evaluating and comparing monocular vision-based approaches and pushing the state-of-the-art on this problem. This work is based on the Satellite Pose Estimation Dataset, the first publicly available machine learning set of synthetic and real spacecraft imageries. The choice of dataset reflects one of the unique challenges associated with spaceborne computer vision tasks, namely the lack of spaceborne images to train and validate the developed algorithms. This work briefly reviews the basic properties and the collection process of the dataset which was made publicly available. The competition design, including the definition of performance metrics and the adopted testbed, is also discussed. The main contribution of this paper is the analysis of the submissions of the 48 competitors, which compares the performance of different approaches and uncovers what factors make the satellite pose estimation problem especially challenging.

cs.CV↗

arXiv2020

Closed-Form Optimal Impulsive Control of Spacecraft Relative Motion Using Reachable Set Theory

This paper addresses the spacecraft relative orbit reconfiguration problem of minimizing the delta-v cost of impulsive control actions while achieving a desired state in fixed time. The problem is posed in relative orbit element (ROE) space, which yields insight into relative motion geometry and allows for the straightforward inclusion of perturbations in linear time-variant form. Reachable set theory is used to translate the cost-minimization problem into a geometric path-planning problem and formulate the reachable delta-v minimum, a new metric to assess optimality and quantify reachability of a maneuver scheme. Next, this paper presents a methodology to compute maneuver schemes that meet this new optimality criteria and achieve a prescribed reconfiguration. Though the methodology is applicable to any linear time-variant system, this paper leverages a state representation in ROE to derive new globally optimal maneuver schemes in orbits of arbitrary eccentricity. The methodology is also used to generate quantifiably sub-optimal solutions when the optimal solutions are unreachable. Further, this paper determines the mathematical impact of uncertainties on achieving the desired end state and provides a geometric visualization of those effects on the reachable set. The proposed algorithms are tested in realistic reconfiguration scenarios and validated in a high-fidelity simulation environment.

eess.SY↗

arXiv2020

Spacecraft Swarm Dynamics and Control About Asteroids

This paper presents a novel methodology to control spacecraft swarms about singleasteroids with arbitrary gravitational potential coefficients. This approach enablesthe use of small, autonomous swarm spacecraft in conjunction with a mothership,reducing the need for the Deep Space Network and increasing safety in future as-teroid missions. The methodology is informed by a semi-analytical model for thespacecraft absolute and relative motion that includes relevant gravitational effectswithout assuming J2-dominance as well as solar radiation pressure. The dynamicsmodel is exploited in an Extended Kalman Filter (EKF) to produce an osculating-to-mean relative orbital element (ROE) conversion that is asteroid agnostic. Theresulting real-time relative mean state estimate is utilized in the formation-keepingcontrol algorithm. The control problem is cast in mean relative orbital elements toleverage the geometric insight of secular and long-period effects in the definitionof control windows for swarm maintenance. Analytical constraints that ensurecollision avoidance and enforce swarm geometry are derived and enforced in ROEspace. The proposed swarm-keeping algorithms are tested and validated in high-fidelity simulations for a reference asteroid mission.

physics.space-ph↗

arXiv2019

Second-Order Solution for Relative Motion on Eccentric Orbits in Curvilinear Coordinates

A new, second-order solution in curvilinear coordinates is introduced for the relative motion of two spacecraft on eccentric orbits. The second-order equations for unperturbed orbits are derived in spherical coordinates with true anomaly as the independent variable, and solved by the method of successive approximations. A comparison of error trends against eccentricity and inter-spacecraft separation is presented between the new solution and prominent Cartesian, curvilinear, and orbital element based solutions from the literature. The second-order curvilinear solution offers a thousand-fold improvement in accuracy over the first-order curvilinear solution, and still greater improvement over first- and second-order rectilinear solutions when large along-track separations are present.

math.DS↗

arXiv2019

Towards Robust Learning-Based Pose Estimation of Noncooperative Spacecraft

This work presents a novel Convolutional Neural Network (CNN) architecture and a training procedure to enable robust and accurate pose estimation of a noncooperative spacecraft. First, a new CNN architecture is introduced that has scored a fourth place in the recent Pose Estimation Challenge hosted by Stanford's Space Rendezvous Laboratory (SLAB) and the Advanced Concepts Team (ACT) of the European Space Agency (ESA). The proposed architecture first detects the object by regressing a 2D bounding box, then a separate network regresses the 2D locations of the known surface keypoints from an image of the target cropped around the detected Region-of-Interest (RoI). In a single-image pose estimation problem, the extracted 2D keypoints can be used in conjunction with corresponding 3D model coordinates to compute relative pose via the Perspective-n-Point (PnP) problem. These keypoint locations have known correspondences to those in the 3D model, since the CNN is trained to predict the corners in a pre-defined order, allowing for bypassing the computationally expensive feature matching processes. This work also introduces and explores the texture randomization to train a CNN for spaceborne applications. Specifically, Neural Style Transfer (NST) is applied to randomize the texture of the spacecraft in synthetically rendered images. It is shown that using the texture-randomized images of spacecraft for training improves the network's performance on spaceborne images without exposure to them during training. It is also shown that when using the texture-randomized spacecraft images during training, regressing 3D bounding box corners leads to better performance on spaceborne images than regressing surface keypoints, as NST inevitably distorts the spacecraft's geometric features to which the surface keypoints have closer relation.

cs.CV↗

arXiv2019

Pose Estimation for Non-Cooperative Rendezvous Using Neural Networks

This work introduces the Spacecraft Pose Network (SPN) for on-board estimation of the pose, i.e., the relative position and attitude, of a known non-cooperative spacecraft using monocular vision. In contrast to other state-of-the-art pose estimation approaches for spaceborne applications, the SPN method does not require the formulation of hand-engineered features and only requires a single grayscale image to determine the pose of the spacecraft relative to the camera. The SPN method uses a Convolutional Neural Network (CNN) with three branches to solve for the pose. The first branch of the CNN bootstraps a state-of-the-art object detector to detect a 2D bounding box around the target spacecraft. The region inside the bounding box is then used by the other two branches of the CNN to determine the attitude by initially classifying the input region into discrete coarse attitude labels before regressing to a finer estimate. The SPN method then uses a novel Gauss-Newton algorithm to estimate the position by using the constraints imposed by the detected 2D bounding box and the estimated attitude. The secondary contribution of this work is the generation of the Spacecraft PosE Estimation Dataset (SPEED). SPEED consists of synthetic as well as actual camera images of a mock-up of the Tango spacecraft from the PRISMA mission. The synthetic images are created by fusing OpenGL-based renderings of the spacecraft's 3D model with actual images of the Earth captured by the Himawari-8 meteorological satellite. The actual camera images are created using a 7 degrees-of-freedom robotic arm, which positions and orients a vision-based sensor with respect to a full-scale mock-up of the Tango spacecraft. The SPN method, trained only on synthetic images, produces degree-level attitude error and cm-level position errors when evaluated on the actual camera images not used during training.

cs.CV↗

arXiv2018

Pose Estimation for Non-Cooperative Spacecraft Rendezvous Using Convolutional Neural Networks

On-board estimation of the pose of an uncooperative target spacecraft is an essential task for future on-orbit servicing and close-proximity formation flying missions. However, two issues hinder reliable on-board monocular vision based pose estimation: robustness to illumination conditions due to a lack of reliable visual features and scarcity of image datasets required for training and benchmarking. To address these two issues, this work details the design and validation of a monocular vision based pose determination architecture for spaceborne applications. The primary contribution to the state-of-the-art of this work is the introduction of a novel pose determination method based on Convolutional Neural Networks (CNN) to provide an initial guess of the pose in real-time on-board. The method involves discretizing the pose space and training the CNN with images corresponding to the resulting pose labels. Since reliable training of the CNN requires massive image datasets and computational resources, the parameters of the CNN must be determined prior to the mission with synthetic imagery. Moreover, reliable training of the CNN requires datasets that appropriately account for noise, color, and illumination characteristics expected in orbit. Therefore, the secondary contribution of this work is the introduction of an image synthesis pipeline, which is tailored to generate high fidelity images of any spacecraft 3D model. The proposed technique is scalable to spacecraft of different structural and physical properties as well as robust to the dynamic illumination conditions of space. Through metrics measuring classification and pose accuracy, it is shown that the presented architecture has desirable robustness and scalable properties.

cs.CV↗