United States Space Force

The detection of dim artificial Earth satellites using ground-based electro-optical sensors, particularly in the presence of background light, is technologically challenging. This perceptual task is foundational to our understanding of the space environment, and grows in importance as the number, variety, and dynamism of space objects increases. We present a hybrid image- and event-based architecture that leverages dynamic vision sensing technology to detect resident space objects in geosynchronous Earth orbit. Given the asynchronous, one-dimensional image data supplied by a dynamic vision sensor, our architecture applies conventional image feature extractors to integrated, two-dimensional frames in conjunction with point-cloud feature extractors, such as PointNet, in order to increase detection performance for dim objects in scenes with high background activity. In addition, an end-to-end event-based imaging simulator is developed to both produce data for model training as well as approximate the optimal sensor parameters for event-based sensing in the context of electrooptical telescope imagery. Experimental results confirm that the inclusion of point-cloud feature extractors increases recall for dim objects in the high-background regime.

View Video Presentation: https://doi.org/10.2514/6.2023-2687.vid In ten years, it will be possible to have an autonomous and persistent small-satellite factory operating in low-earth orbit. This is enabled by advances in technologies over the past two decades: low-cost launch, additive manufacturing, robotics, and artificial intelligence. By developing a roadmap to achieve this end goal, it was possible to identify research topics and prioritize near-term low-cost research efforts. To facilitate the roadmap, an orbital satellite factory concept has been developed that combines proven orbital manufacturing techniques with other technologies that could be demonstrated on orbit within several years. The factory concept relies on robotics and specialized manufacturing tools to create satellite structures with integrated electronics and assembly of complicated parts. With the conceptual design complete, the research required to bring that concept (or one like it) to fruition becomes more evident. Recommended research topics and key milestones for the next five years are provided, from which a few are being pursued actively. Starting now is crucial since the time to progress from terrestrial development through demonstration in vacuum to orbital testing will take several years, and the efforts would be difficult to accelerate due to the nature of the research. The early steps have the lowest cost, yet they have the greatest impact in the development timescale. The next step will be the finalization of the roadmap, identification of research, and identification of parties across industry, government, and academia to help fund and pursue the research.

Star localization in astronomical imagery is a computer vision task that underpins satellite tracking. Astronomical star extraction techniques often struggle to detect stars when applied to satellite tracking imagery due to the narrower fields of view and rate track observational modes of satellite tracking telescopes. We present a large dataset of real narrow-field rate-tracked imagery with ground truth stars, created using a combination of existing star detection techniques, an astrometric engine, and a star catalog. We train three state of the art object detection, instance segmentation, and line segment detection models on this dataset and evaluate them with object-wise, pixel-wise, and astrometric metrics. Our proposed approaches require no metadata; when paired with a lost-in-space astrometric engine, they find astrometric fits based solely on uncorrected image pixels. Experimental results on real data indicate the effectiveness of learned star detection: we report astrometric fit rates over double that of classical star detection algorithms, improved dim star recall, and comparable star localization residuals.

The application of machine learning to a task often necessitates the production of synthetic training data. Some tasks involve rare, but important, scenarios that may not yet have been observed; others are difficult to collect or annotate in large volumes. These difficulties are particularly acute in computer vision applications to scientific imagery, in which human annotation is complicated by noise, ambiguity, and interpretation. One such application is the detection of resident space objects (RSOs) in electro-optical images for space domain awareness (SDA). In many cases, the mislabeling of RSOs by an imperfect annotator (human or machine) can be detrimental to machine learning model performance, especially when the signal-to-noise (SNR) is near or below human detection levels. In this work we introduce SatSim, a modular electro-optical synthetic data generation engine designed to procedurally generate representative, annotated synthetic electro-optical imagery of remote space scenes. SatSim enables rapid generation of synthetic data through Graphics Processing Unit (GPU) acceleration with TensorFlow. This paper discusses the use of SatSim to enhance machine learning approaches and reports the performance of models trained with real data, synthetic data, and real data augmented with synthetic RSOs. In addition, we explore using SatSim to evaluate current state-of-the-art RSO detection algorithms with new sensors (such as all-sky and event-based) and rare but critically important scenarios (such as satellite breakups and collisions) for which limited real data are available.

View Video Presentation: https://doi.org/10.2514/6.2021-4219.vid The interagency Space Science and Technology Partnership Forum was established in 2015 to identify synergistic efforts and technologies across the U.S. government. While the various space agencies of the U.S. government have distinctly different visions for future operational space systems, all share important foundational common needs. These needs, combined with the maturation of autonomous technology and the prospect of leveraging autonomous systems to address those needs, have led each agency to consider how and when to implement increasing levels of autonomy in their space systems, and how to determine the trustworthiness of an autonomous system. The Partnership facilitated dialogue among the partners, collected and analyzed data on current and desired future levels of capability, and identified gaps to motivate three recommendations that can be addressed within the Partnership community. These recommendations address the need for more robust documenting and socializing of anomalies in space system operations; the need to expand communication and trust within the community of developers, operators, and end users; and the need for a safe development and testing environment for maturing and demonstrating future autonomous space systems. These recommendations will facilitate both near-term programmatic actions and long-term steps for implementing enduring progress towards enabling space trusted autonomy.

Effective space domain awareness requires positive identification of artificial satellites. Current methods for extracting object identification from observed data require spatially resolved imagery which limits identification to objects in low earth orbits. Many artificial Earth satellites, however, operate in geostationary orbits at distances which prohibit ground based observatories from resolving spatial information. This paper demonstrates an object identification solution leveraging modified residual convolutional neural networks to map distance-invariant spectroscopic data to object identity. We report classification accuracies exceeding 80% for a simulated 64-class satellite problem−even in the case of satellites undergoing constant, random re-orientation. An astronomical observing campaign driven by these results returned accuracies of ∼72% for a nine-class problem with an average of 100 examples per class, performing as expected from simulation. We demonstrate the application of variational Bayesian inference by dropout, stochastic weight averaging (SWA), and SWA-focused deep ensembling to measure classification uncertainties−critical components in space domain awareness where routine decisions risk expensive space assets and carry geopolitical consequences.

The KC-135R Stratotanker is a multifunction slim body aircraft that provides air refueling and airlift for the United States’ war and peacetime requirements, which demand a certain level of availability. As the KC-135R fleet ages, the aircraft availability (AA) rate degrades due to high demand use, stress, and the age of the equipment. Preventative and corrective maintenance is designed to return the aircraft to an available state to meet mission requirements, but the United States Air Force continues to fail at meeting every requirement communicated by commanders for KC-135R air refueling and airlift. Focusing on aircraft metrics can enable a prediction model of AA and provide the unit commanders the tools for data influenced decisions. The analysis of historical aircraft maintenance and flight metrics will show a correlation between tracked metrics and AA, thus, the ability to predict a future availability rate. Furthermore, analyzing the data with machine learning techniques will improve prediction accuracy by evaluating the variable importance and will make inferences learned from mining the data that are otherwise difficult to model in a complex system of systems.

View Video Presentation: https://doi.org/10.2514/6.2023-1837.vid In this paper, the applications of ceramic matrix composites (CMCs) in liquid rocket engines (LRE) are investigated from the perspective of several years’ experience with volume production of hot-section CMC components for air-breathing gas turbine engines. SiC/SiC CMC applications are proposed that offer the best advantages, and immediate developments needed to facilitate the application of CMCs in LREs are identified. The technology readiness, characteristics, advantages, and disadvantages of CMCs in liquid rocket engines are discussed. A performance model of a generic gas generator cycle rocket engine indicates that engine Isp could increase by up to 5.5 seconds if the turbine inlet temperature could be raised to 2,200 K, a capability that may be enabled with CMCs. For a typical mission consisting of a 5-minute burn operating at 890 kN thrust, this results in a reduction of consumed propellant of 2,040 kg. A separate thermostructural analytical study was conducted using a representative regenerative cooling channel made of CMC subject to realistic environments. Results indicate that CMCs may outperform traditional materials given the expected range of conditions, particularly over repeated cycles, and at significantly lower weight.

Here we present a concept for a mobile, completely off-grid, robotic observatory for rapid deployment and observational support. This 1-meter aperture, 3-degree FOV telescope employs state-of-the-art commercial instrumentation such that it not only supports satellite orbit cataloging but also closely spaced object detection/ characterization at atmospheric seeing limits, i.e. sub-arcsecond pixels vs. more traditional cataloging systems’ 2-3 arcsecond pixels. Its relatively large étendue, high throughput, and up to 50 deg/s slew speeds provides for high survey speeds, be it for lost space debris or astronomical transients. We will detail the design and simulated performance of this Deployable, Attritable Optical (DAO) system. Furthermore, each system will employ US Space Force developed observatory control software called SensorKit, a completely open-source software that enables robotic operation and, if desired for SDA purposes, communication with the Unified Data Library. Scheduling, tasking, data processing and dissemination and more are a part of the US Space Force MACHINA program, presented separately in these proceedings.

The U.S. Space Force’s Space and Missile Systems Center’s Launch Enterprise and The Aerospace Corporation (Launch Enterprise/Aerospace) continue to advance the systems engineering, mission assurance, and acquisition processes for National Security Space Launch (NSSL) services. The Launch Enterprise/Aerospace team has proactively evolved the mission assurance process to embrace new capabilities such as reusable launch vehicles and additive manufacturing. The mission assurance process also evolved to incorporate new technologies, increase flexibility, reduce cost, and incorporate lessons learned for application to expendable launch vehicles and reusable launch systems. This paper provides a top-level summary on 1) the advancements made to the mission assurance process, 2) the development of new technical standards, 3) the tailoring of standards, and 4) the adoption of standards to address lessons learned. Furthermore, the application of these advances to the NSSL acquisition of launch services is also discussed because it is a key U.S. Space Force objective to launch every National Security Space satellite with high reliability and low risk to achieve full operational capability.

The Space Development Agency (SDA) is developing the Proliferated Warfare Space Architecture (PWSA) – a constellation of hundreds of satellites in low earth orbit delivering space-based capabilities to the joint warfighter. The PWSA is a mesh network of optically connected satellites providing low-latency data transport and missile warning/tracking capabilities. SDA capitalizes on a unique business model that values speed and lowers costs by harnessing commercial development. The Optical Communications Terminal (OCT) standard was created to provide optical interoperability specifications, enable a strong marketplace, and to drive advancements in optical communication capabilities to terrestrial, maritime, and airborne warfighting elements. As part of the spiral development process, the OCT standard evolves with PWSA deployment phases. SDA has incorporated feedback as well as advancements to the OCT standard, resulting in the release of version 3.1.0. In this paper we discuss key aspects of the OCT standard, such as wavelength, modulation, data rates, polarization, link distance, error correction coding, pointing, acquisition and tracking, and position, navigation, and timing.

Ground-based imaging of objects in Low Earth Orbit (LEO) is complicated by atmospheric turbulence, which make it difficult to identify key features or components on the object of interest. Many automated image reconstruction techniques are in use, but expert labor is needed to subjectively discern and identify truth features on a partially reconstructed image. In this paper, we present a deep learning approach for semantic segmentation of ground-based images of LEO objects. We investigate the performance under various atmospheric turbulence strengths in terms of the Fried parameter (<tex>$r_{0}$</tex>) and show the viability of this method.

Recent work demonstrates recognition of artificial satellites in spatially unresolved observations by utilizing learned spectroscopic classification (SpectraNet1 ). That proof of concept exposes critical identifying information currently lacking in catalogs used by space domain awareness stakeholders. In this work we present experiments to increase the accessibility and efficiency of SpectraNet enabled systems by probing the bandpass and resolution requirements for learned recognition of satellites. To enable affordable, off the shelf instrumentation, this work focuses on wavelength ranges accessible by Silicon-based detectors (400-1000 nanometers). While the SpectraNet proof of concept utilized a medium resolution spectrograph on a 3.6 meter telescope at 10,000 feet elevation, we show that the identifying spectral features relate to an object’s overall spectral energy density and are accessible at significantly lower spectral resolution. This finding relaxes the need for large telescopes at high altitude. We further demonstrate that the technology can be utilized via simultaneous multi-band filter photometry. Design considerations for properly obtaining simultaneous photometry are discussed. Thus this work demonstrates that−in simulation−learned spectral recognition is an effective technology from high resolution spectrographs through simultaneous multi-filter photometric instruments. We provide experiments to understand the minimum engineered system needed to perform effective learned recognition, such that the technology can be hardened and widely proliferated.

We introduce two new tools to the application of polarimetry to space domain awareness (SDA), the LoVIS spectropolarimeter on the 3.6 m AEOS telescope and deep convolutional neural networks (CNNs). Using a dataset of 20,000 simulated satellite observations, we train a CNN to map distance-invariant spectropolarimetric data to object identity. We report the classification accuracy of this simulation for a 9-class satellite problem, comparing results against low-resolution spectra for which prior success has been demonstrated as well as solar phase angle and satellite apparent magnitude. These initial experiments show potential for improved discrimination against nearly identical satellites on the basis of added polarimetric data.

View Video Presentation: https://doi.org/10.2514/6.2023-2214.vid To overcome the technical readiness levels (TRL) 4 – 7 “Valley of Death” and accelerate new technology development, the US Space Force (USSF) started an in-Space Developmental Test (iSDT) initiative with its instantiation of a persistent iSDT platform called the Advanced Space-Based Testbed (XST). For risk-reduction, a scaled-down Mini-XST is proposed for proof-of-concept. Its key space systems include the spacecraft bus, test payloads, tender servicer to deliver test payload to the Mini-XST spacecraft, and robotic servicer to attach test payloads to the Mini-XST platform. In this paper, the proposed Mini-XST taking advantage of recent proliferated small satellite technologies as well as standard interface to support its successful operation will be discussed and illustrated.

In this report, the authors present findings addressing the potential use of live, virtual, and constructive (LVC) simulation capabilities for continuation training at air operation centers in the U.S. Air Force, in support of Joint All Domain Command and Control (JADC2). Training to support JADC2 requires preemptive consideration of supporting capabilities that align with training needs. LVC capabilities can help support this complex training.

The detection of closely spaced artificial satellites informs tactical decision making in a high risk scenario in the space domain. In regimes where spatial information is lost (ground observations of small or distant satellites), spectroastrometry simulations have demonstrated the potential to detect the presence of multiple objects down to 0′′.05–ten meters at geostationary orbit–using a medium resolution optical spectrograph on a large aperture telescope.¹ This technique falls into the growing field of<i> learned</i> space domain awareness: leveraging convolutional neural networks to rapidly infer tactical information from complex, non-intuitive data. In this work we present a field rotation nodding technique that removes the need for a priori knowledge of the closely spaced object on sky orientation. We discuss modifications to an optical spectrograph necessary to perform this technique. We present simulated bounds on the effectiveness of spectroastrometry for the detection of closely spaced objects.

Multi-frame blind deconvolution (MFBD) algorithms are able to produce high-resolution image reconstructions from severely degraded inputs. Often these algorithms are designed with a number of assumptions about the observing scenario and the data quality, and when these assumptions are violated the reconstruction quality can suffer. However, it can be challenging to automatically assign quality scores to input images for data rejection, especially while observing Low Earth Orbit (LEO) satellites through the turbulent atmosphere as they transit across a large patch of the sky. We report on an algorithm that uses a convolutional neural network (CNN) to assign quality scores to images prior to MFBD processing in a system titled Quality-Weighted Iterative Deconvolution (QWID). This quality assessment represents the likelihood that each input frame can contribute meaningful signal to the frame reconstruction process. The neural network is trained on a simulated dataset of ground-based observations of LEO satellites, where true quality is known. Improvements in performance over the same MFBD implementation in the absence of quality scores are demonstrated on a subset of these simulated observations.

This paper discusses the 2030 Space Logistics vision, and the work done by the USSF Pervasive Capability Collaboration Teams (CCTs) and the Space Access, Mobility & Logistics (SAML) CCT to support in-space Servicing, Mobility, & Logistics (SML) requirements; these in-space capabilities include in-space autonomous rendezvous, proximity operations, and docking (RPOD), active debris mitigation and disposal of non-cooperative (tumbling) space objects, small, efficient, low-cost, trusted, autonomous robotic arms, and open standard interfaces to enable ISAM, and SML technical need solutions and transition. This paper will also propose more work to be done to support SML in the future, including modeling in-space logistics as a supply-aggregation-and-distribution network utilizing a hub-and-spoke (node-and-leg) framework, and architecture-level interoperability issues.

In this work, we propose to consolidate radio frequency communication signals and speech audio into a common data modality: multichannel, time-continuous amplitudes with characteristic spectrograms and a finite symbol alphabet. By putting a portion of the radio spectrum on a similar footing to audio, this may allow us to leverage a great deal of the technological progress achieved by automatic speech recognition (ASR) and readily transfer it to radio frequency machine learning (RFML), a rapidly developing field. To support this claim, we take the leading ASR architecture of wav2vec2 and apply it directly to a challenging dataset of real, low-SNR radio signals captured from satellite telecommunications. Representing the first large-scale application of learned detection and classification of raw signals emitted from a diverse array of active low Earth orbit satellites, the speech-inspired network demonstrates strong proficiency on all tasks and robustness to the degraded signal environment.