SEARCH · Engineering Papers

Results for “random networks”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 271 records · Page 15

All sources NASA NTRS DOE OSTI

NASA NTRSDate not supplied

Machine Learning Models to Predict Cognitive Impairment of Rodents Subjected to Space Radiation

INTRODUCTION We use artificial neural networks (ANNs) as an example machine learning (ML) tool to predict the cognitive performance impairment of rats induced by irradiation. The experimental data in the analyses is attentional set-shifting (ATSET) test scores from a rodent model exposed to ≤15 cGy of individual galactic cosmic radiation (GCR) ions: 4He, 28Si, or 56Fe, expected for a Lunar or Mars mission [1]. This work investigates rats at a subject-based level and uses applied dose and performance scores taken before irradiation to predict whether a rat will be impaired when irradiated. The results of this study are significant to crewed space missions as they support the potential of predicting an astronaut’s impairment in a specific task before spaceflight through the implementation of appropriately trained ML tools. METHODS Data used in this work are scores from the ATSET, a multi-stage constrained cognitive flexibility test [2]. Our computational model utilizes the number of attempts to reach the criterion to pass a stage as a behavioral performance measure for rats. We use the post-irradiation scores, generate thresholds from cumulative distribution plots of non-irradiated rats, and calculate the percent of irradiated rats whose scores fall below the threshold to infer how each radiation type/dose affects a population. Rats scoring above the threshold are labeled impaired while the others are non-impaired. We then employ ANNs as a typical ML technique, and use each subject’s individual scores taken before radiation along with the applied dose, to predict their personal susceptibility to cognitive impairment due to space radiation exposure. RESULTS AND CONCLUSION A significant finding is the exhibition of a dose-dependent increasing probability of impairment for 1 to 10 cGy of 28Si or 56Fe in the simple discrimination (SD) stage of the ATSET, and for 1 to 10 cGy of 56Fe in the compound discrimination (CD) stage. On a subject-based level, implementing ML classifiers such as ANNs identifies rats that have a higher tendency for impairment after GCR exposure [1]. The receiver operating characteristic (ROC) and the precision-recall (PR) curves of the ML models show a better prediction of impairment when 56Fe is the ion in question in both SD (Figure 1) and CD stages. They, however, do not depict impairment due to 4He in SD (Figure 1) and 28Si in CD, suggesting no dose-dependent impairment response in these cases. In this work, “good” prediction pertains to “better-than-random-chance”, due to the limited sample size and the high inter- and intra-individual variabilities in response to brain stimulation paradigms, as applicable to both animals and humans. More behavioral tests and biomarkers should be investigated on the same subjects, to be fed to the ML models to capture the agents responsible for performance alterations of some individuals versus others.

machine learning↗

NASA NTRS1987

A ray tracing model for leaf bidirectional scattering studies

A leaf is modeled as a deterministic two-dimensional structure consisting of a network of circular arcs designed to represent the internal morphology of major species. The path of an individual ray through the leaf is computed using geometric optics. At each intersection of the ray with an arc, the specular reflected and transmitted rays are calculated according to the Snell and Fresnel equations. Diffuse scattering is treated according to Lambert's law. Absorption is also permitted but requires a detailed knowledge of the spectral attenuation coefficients. An ensemble of initial rays are chosen for each incident direction with the initial intersection points on the leaf surface selected randomly. The final equilibrium state after all interactions then yields the leaf bidirectional reflectance and transmittance distributions. The model also yields the internal two dimensional light gradient profile of the leaf.

Brakke, T. W.↗

NASA NTRS2020

Amino Acid Encoding for Deep Learning Applications

Background: The number of applications of deep learning algorithms in bioinformatics is increasing as they usually achieve superior performance over classical approaches, especially, when bigger training datasets are available. In deep learning applications, discrete data, e.g. words or n-grams in language, or amino acids or nucleotides in bioinformatics, are generally represented as a continuous vector through an embedding matrix. Recently, learning this embedding matrix directly from the data as part of the continuous iteration of the model to optimize the target prediction – a process called ‘end-to-end learning’ – has led to state-ofthe-art results in many fields. Although usage of embeddings is well described in the bioinformatics literature, the potential of end-to-end learning for single amino acids, as compared to more classical manually-curated encoding strategies, has not been systematically addressed. To this end, we compared classical encoding matrices, namely one-hot, VHSE8 and BLOSUM62, to end-to-end learning of amino acid embeddings for two different prediction tasks using three widely used architectures, namely recurrent neural networks (RNN), convolutional neural networks (CNN), and the hybrid CNN-RNN. Results: By using different deep learning architectures, we show that end-to-end learning is on par with classical encodings for embeddings of the same dimension even when limited training data is available, and might allow for a reduction in the embedding dimension without performance loss, which is critical when deploying the models to devices with limited computational capacities. We found that the embedding dimension is a major factor in controlling the model performance. Surprisingly, we observed that deep learning models are capable of learning from random vectors of appropriate dimension. Conclusion: Our study shows that end-to-end learning is a flexible and powerful method for amino acid encoding. Further, due to the flexibility of deep learning systems, amino acid encoding schemes should be benchmarked against random vectors of the same dimension to disentangle the information content provided by the encoding scheme from the distinguishability effect provided by the scheme.

Hesham ElAbd↗

NASA NTRS1999

Using a Genetic Algorithm to Model Broadband Regional Waveforms for Crustal Structure in the Western United States

In this study, we analyze regional seismograms to obtain the crustal structure in the eastern Great Basin and western Colorado plateau. Adopting a for- ward-modeling approach, we develop a genetic algorithm (GA) based parameter search technique to constrain the one-dimensional crustal structure in these regions. The data are broadband three-component seismograms recorded at the 1994-95 IRIS PASSCAL Colorado Plateau to Great Basin experiment (CPGB) stations and supplemented by data from U.S. National Seismic Network (USNSN) stations in Utah and Nevada. We use the southwestern Wyoming mine collapse event (M(sub b) = 5.2) that occurred on 3 February 1995 as the seismic source. We model the regional seismograms using a four-layer crustal model with constant layer parameters. Timing of teleseismic receiver functions at CPGB stations are added as an additional constraint in the modeling. GA allows us to efficiently search the model space. A carefully chosen fitness function and a windowing scheme are added to the algorithm to prevent search stagnation. The technique is tested with synthetic data, both with and without random Gaussian noise added to it. Several separate model searches are carried out to estimate the variability of the model parameters. The average Colorado plateau crustal structure is characterized by a 40-km-thick crust with velocity increases at depths of about 10 and 25 km and a fast lower crust while the Great Basin has approximately 35- km-thick crust and a 2.9-km-thick sedimentary layer.

Bhattacharyya, Joydeep↗

NASA NTRS1987

Power supply conditioning circuit

A power supply conditioning circuit that can reduce Periodic and Random Deviations (PARD) on the output voltages of dc power supplies to -150 dBV from dc to several KHz with no measurable periodic deviations is described. The PARD for a typical commercial low noise power supply is -74 dBV for frequencies above 20 Hz and is often much worse at frequencies below 20 Hz. The power supply conditioning circuit described here relies on the large differences in the dynamic impedances of a constant current diode and a zener diode to establish a dc voltage with low PARD. Power supplies with low PARD are especially important in circuitry involving ultrastable frequencies for the Deep Space Network.

Primas, L. E.↗

NASA NTRS2010

Initialization and Restart in Stochastic Local Search: Computing a Most Probable Explanation in Bayesian Networks

For hard computational problems, stochastic local search has proven to be a competitive approach to finding optimal or approximately optimal problem solutions. Two key research questions for stochastic local search algorithms are: Which algorithms are effective for initialization? When should the search process be restarted? In the present work we investigate these research questions in the context of approximate computation of most probable explanations (MPEs) in Bayesian networks (BNs). We introduce a novel approach, based on the Viterbi algorithm, to explanation initialization in BNs. While the Viterbi algorithm works on sequences and trees, our approach works on BNs with arbitrary topologies. We also give a novel formalization of stochastic local search, with focus on initialization and restart, using probability theory and mixture models. Experimentally, we apply our methods to the problem of MPE computation, using a stochastic local search algorithm known as Stochastic Greedy Search. By carefully optimizing both initialization and restart, we reduce the MPE search time for application BNs by several orders of magnitude compared to using uniform at random initialization without restart. On several BNs from applications, the performance of Stochastic Greedy Search is competitive with clique tree clustering, a state-of-the-art exact algorithm used for MPE computation in BNs.

Mengshoel, Ole J.↗

NASA NTRSDate not supplied

Optimizing Altitude Sampling and Sensitivity with the Goldstone Orbital Debris Radar

The NASA Orbital Debris Program Office (ODPO) has used the Goldstone Orbital Debris Radar (Goldstone) since 1993 to characterize orbital debris (OD) in low Earth orbit too small to be tracked by the U.S. Space Surveillance Network. Operated by NASA’s Jet Propulsion Laboratory, Goldstone can measure OD as small as 3 mm at 1000 km altitude and lower. Goldstone is a bistatic radar that for 25 years used Deep Space Station (DSS)-14 as a transmitter and DSS-15 as a receiver. In early 2018, DSS-15 was decommissioned and replaced with DSS-25 (and occasionally DSS-26) of the Deep Space Network Apollo Cluster. The increased baseline between DSS-14 and DSS-25 significantly reduced the instantaneous altitude coverage of the bistatic beam overlap. Initial measurements in 2018 were focused around 800 km, which has approximately the highest flux of sub-centimeter debris. In 2019, DSS-14 was offline for maintenance, and the ODPO designed an annual survey observation plan to efficiently sample altitudes from 700 km to 1000 km, since many NASA satellites fly in this range. This paper discusses the observation plan, including the development of the pointings, a refinement of the altitudes of interest, and an analysis of the effects of random pointing errors on beam overlap. Additionally, results from measurements taken in 2020 and 2021 are presented, showing that not only is the observation plan effective at sampling 700 km to 1000 km altitude, but it is also producing the most sensitive terrestrial radar measurements at these altitudes to date.

James Murray↗

NASA NTRS2023

Optimizing Altitude Sampling and Sensitivity with the Goldstone Orbital Debris Radar

James Murray↗

NASA NTRS2001

The Use of CASES-97 Observations to Assess and Parameterize the Impact of Land-Surface Heterogeneity on Area-Average Surface Heat Fluxes for Large-Scale Coupled Atmosphere-Hydrology Models

To understand the effects of land-surface heterogeneity and the interactions between the land-surface and the planetary boundary layer at different scales, we develop a multiscale data set. This data set, based on the Cooperative Atmosphere-Surface Exchange Study (CASES97) observations, includes atmospheric, surface, and sub-surface observations obtained from a dense observation network covering a large region on the order of 100 km. We use this data set to drive three land-surface models (LSMs) to generate multi-scale (with three resolutions of 1, 5, and 10 kilometers) gridded surface heat flux maps for the CASES area. Upon validating these flux maps with measurements from surface station and aircraft, we utilize them to investigate several approaches for estimating the area-integrated surface heat flux for the CASES97 domain of 71x74 square kilometers, which is crucial for land surface model development/validation and area water and energy budget studies. This research is aimed at understanding the relative contribution of random turbulence versus organized mesoscale circulations to the area-integrated surface flux at the scale of 100 kilometers, and identifying the most important effective parameters for characterizing the subgrid-scale variability for large-scale atmosphere-hydrology models.

Chen, Fei↗

NASA NTRS2021

Design and Modeling of a Coded-Interleaving System in the Presence of Fading

Future deep space links are migrating towards higher frequency bands such as Ka band and optical. These links are susceptible to non linear effects (e.g. cloud coverage, atmospheric turbulence, scintillation, antenna mis-pointing, jitter, relative motion, etc.) that cause various degrees of fading loss over various time scales. We consider a channel experiencing block fading where the channel coherence time is a lot longer than the length of a codeword. As long as the fading duration is not excessive, an interleaver can be used in a coded-communication system that randomizes the fades among symbols of multiple codewords. This achieves an “averaging” effect that spreads the signal energy more evenly among the codewords, and improves the likelihood that each of them can be decoded correctly. For simplicity, we consider a block interleaver where adjacent symbols of a codeword have a fixed delay (or separation) with each other, and investigate the error-rate performance of a coded-interleaving system (with fixed and finite interleaving depth) in the presence of block fading. Unlike prior research which assumes theoretical channel coherence models, we aim to build upon our analysis using empirical spacecraft signal measurements obtained with the Deep Space Network, and to use the fundamental theories of probability and statistics (e.g. variations of Law of Large Numbers) to quantify the error-rate performance of the coded-interleaving system. We propose to utilize these results in the design of a two-dimensional coded-interleaving system for fading channels.

Sanchez-Net, Marc↗

NASA NTRS1998

Automatic Molecular Design using Evolutionary Techniques

Molecular nanotechnology is the precise, three-dimensional control of materials and devices at the atomic scale. An important part of nanotechnology is the design of molecules for specific purposes. This paper describes early results using genetic software techniques to automatically design molecules under the control of a fitness function. The fitness function must be capable of determining which of two arbitrary molecules is better for a specific task. The software begins by generating a population of random molecules. The population is then evolved towards greater fitness by randomly combining parts of the better individuals to create new molecules. These new molecules then replace some of the worst molecules in the population. The unique aspect of our approach is that we apply genetic crossover to molecules represented by graphs, i.e., sets of atoms and the bonds that connect them. We present evidence suggesting that crossover alone, operating on graphs, can evolve any possible molecule given an appropriate fitness function and a population containing both rings and chains. Prior work evolved strings or trees that were subsequently processed to generate molecular graphs. In principle, genetic graph software should be able to evolve other graph representable systems such as circuits, transportation networks, metabolic pathways, computer networks, etc.

Globus, Al↗

NASA NTRS2023

A Hyperspectral Inversion Framework for Estimating Absorbing Inherent Optical Properties and Biogeochemical Parameters in Inland and Coastal Waters

The simultaneous remote estimation of biogeochemical parameters (BPs) and inherent optical properties (IOPs) from hyperspectral satellite imagery of globally distributed optically distinct inland and coastal waters is a complex, unsolved, non-unique inverse problem. To tackle this problem, we leverage a machine-learning model termed Mixture Density Networks (MDNs). MDNs outperform operational algorithms by calculating the covariance between the simultaneously estimated products. We train the MDNs on a large ( N = 8237) dataset of co-aligned, in situ measured, hyperspectral remote sensing reflectance (R rs ), BPs, and absorbing IOPs from globally representative optically distinct inland and coastal waters. The estimated IOPs include absorption due to phytoplankton (a ph ), chromophoric dissolved organic matter (a cdom ), and non-algal particles (a nap ). The estimated BPs include chlorophyll-a, total suspended solids, and phycocyanin (PC). MDNs dramatically reduce uncertainty in the retrievals, relative to operational algorithms, when using a 50/50 dataset split, where the MDNs are trained on a randomly selected half of the in situ dataset and validated on the other half. Our model is shown to have higher, or equivalent, generalization performance than the calculated operational algorithms available for all BPs and IOPs (except PC) via a leave-one-out cross-validation assessment. The MDNs are sensitive to uncertainties in the hyperspectral satellite R rs , resulting from instrument noise and atmospheric correction; there is a difference of ~37.4–62.8% (using median symmetric accuracy) between the MDNs’ estimates derived from co-located satellite-derived R rs and in situ R rs . Of the IOPs, a cdom and a nap are less sensitive to uncertainties in hyperspectral satellite imagery relative to a ph , with remote estimates of a ph exhibiting incorrect spectral shape and magnitude relative to in situ measured IOPs. Despite the uncertainties in satellite derived R rs , the spatial distributions of BPs and IOPs in MDN-derived product maps of Lake Erie and the Curonian Lagoon, based on imagery taken with the Hyperspectral Imager for the Coastal Ocean (HICO) and PRecursore Iper-Spettrale della Missione Applicativa (PRISMA), are confirmed via co-aligned in situ measurements and agree with the literature’s understanding of these well-studied regions. The consistency and accuracy of the model on HICO and PRISMA imagery, despite radiometric uncertainties, demonstrate its applicability to future hyperspectral missions, such as the Plankton, Aerosol, Cloud, ocean Ecosystem (PACE) mission, where the simultaneous estimation model will serve as a key part of phytoplankton community composition analysis.

Ryan E. O'Shea↗

NASA NTRS2008

Safe Upper-Bounds Inference of Energy Consumption for Java Bytecode Applications

Many space applications such as sensor networks, on-board satellite-based platforms, on-board vehicle monitoring systems, etc. handle large amounts of data and analysis of such data is often critical for the scientific mission. Transmitting such large amounts of data to the remote control station for analysis is usually too expensive for time-critical applications. Instead, modern space applications are increasingly relying on autonomous on-board data analysis. All these applications face many resource constraints. A key requirement is to minimize energy consumption. Several approaches have been developed for estimating the energy consumption of such applications (e.g. [3, 1]) based on measuring actual consumption at run-time for large sets of random inputs. However, this approach has the limitation that it is in general not possible to cover all possible inputs. Using formal techniques offers the potential for inferring safe energy consumption bounds, thus being specially interesting for space exploration and safety-critical systems. We have proposed and implemented a general frame- work for resource usage analysis of Java bytecode [2]. The user defines a set of resource(s) of interest to be tracked and some annotations that describe the cost of some elementary elements of the program for those resources. These values can be constants or, more generally, functions of the input data sizes. The analysis then statically derives an upper bound on the amount of those resources that the program as a whole will consume or provide, also as functions of the input data sizes. This article develops a novel application of the analysis of [2] to inferring safe upper bounds on the energy consumption of Java bytecode applications. We first use a resource model that describes the cost of each bytecode instruction in terms of the joules it consumes. With this resource model, we then generate energy consumption cost relations, which are then used to infer safe upper bounds. How energy consumption for each bytecode instruction is measured is beyond the scope of this paper. Instead, this paper is about how to infer safe energy consumption estimations assuming that those energy consumption costs are provided. For concreteness, we use a simplified version of an existing resource model [1] in which an energy consumption cost for individual Java opcodes is defined.

Navas, Jorge↗

NASA NTRS1993

AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics, and Materials Conference, 34th and AIAA/ASME Adaptive Structures Forum, La Jolla, CA, Apr. 19-22, 1993, Technical Papers. Pts. 1-6

Topics addressed include the prediction of helicopter component loads using neural networks, spacecraft on-orbit coupled loads analysis, hypersonic flutter of a curved shallow panel with aerodynamic heating, thermal-acoustic fatigue of ceramic matrix composite materials, transition elements based on transfinite interpolation, damage progression in stiffened composite panels, a direct treatment of min-max dynamic response optimization problems, and sources of helicopter rotor hub inplane shears. Also discussed are dynamics of a layered elastic system, confidence bounds on structural reliability, mixed triangular space-time finite elements, advanced transparency development for USAF aircraft, a low-velocity impact on a graphite/PEEK, an automated mode-tracking strategy, transonic flutter suppression by a passive flap, a nonlinear response of composite panels to random excitation, an optimal placement of elastic supports on a simply supported plate, a probabilistic assessment of composite structures, a model for mode I failure of laminated composites, a residual flexibility approach to multibody dynamics,and multilayer piezoelectric actuators.

Source record↗

NASA NTRS2021

Lila: Optimal Dispatching in Probabilistic Temporal Networks using Monte Carlo Tree Search

Executing a Probabilistic Simple Temporal Network (PSTN) amounts at scheduling, i.e. \textit{dispatch}, a set of events under time uncertainty. This constitutes a NP-hard online optimization problem. The right execution time must be dynamically assigned to each event of the PSTN such that the temporal constraints are met, whereas activity durations are progressively observed as the execution unfolds. We propose a dispatching algorithm based on Monte Carlo Tree Search, called Lila, with the following characteristics: (i) it is an anytime algorithm, both offline and online, proven asymptotically optimal; (ii) it returns the current probability of success, either before or at any moment during operations; (iii) it handles any possible continuous or discrete, even non-parametric, probability distributions, as well as inter-dependencies between random variables, exogenous and endogenous uncertainty; and (iv) can be easily extended to handle probabilistic external events, PSTNs with resources, PSTNs with cutoff times and precondition chains, etc. Lila is universal in the sense that it can handle any dispatching protocol, simply by specifying it to the algorithm. It has the unlimited flexibility offered by the simulation paradigm, whilst it asymptotically converges to optimal decisions and/or robustness approximations.

Chien, Steve A.↗

NASA NTRS2010

Analysis of Snow Bidirectional Reflectance from ARCTAS Spring-2008 Campaign

The spring 2008 Arctic Research of the Composition of the Troposphere from Aircraft and Satellites (ARCTAS) experiment was one of major intensive field campaigns of the International Polar Year aimed at detailed characterization of atmospheric physical and chemical processes in the Arctic region. A part of this campaign was a unique snow bidirectional reflectance experiment on the NASA P-3B aircraft conducted on 7 and 15 April by the Cloud Absorption Radiometer (CAR) jointly with airborne Ames Airborne Tracking Sunphotometer (AATS) and ground-based Aerosol Robotic Network (AERONET) sunphotometers. The CAR data were atmospherically corrected to derive snow bidirectional reflectance at high 1 degree angular resolution in view zenith and azimuthal angles along with surface albedo. The derived albedo was generally in good agreement with ground albedo measurements collected on 15 April. The CAR snow bidirectional reflectance factor (BRF) was used to study the accuracy of analytical Ross-Thick Li-Sparse (RTLS), Modified Rahman-Pinty-Verstraete (MRPV) and Asymptotic Analytical Radiative Transfer (AART) BRF models. Except for the glint region (azimuthal angles phi less than 40 degrees), the best fit MRPV and RTLS models fit snow BRF to within 0.05. The plane-parallel radiative transfer (PPRT) solution was also analyzed with the models of spheres, spheroids, randomly oriented fractal crystals, and with a synthetic phase function. The latter merged the model of spheroids for the forward scattering angles with the fractal model in the backscattering direction. The PPRT solution with synthetic phase function provided the best fit to measured BRF in the full range of angles. Regardless of the snow grain shape, the PPRT model significantly over-/underestimated snow BRF in the glint/backscattering regions, respectively, which agrees with other studies. To improve agreement with experiment, we introduced a model of macroscopic snow surface roughness by averaging the PPRT solution over the slope distribution function and by adding a simple model of shadows. With macroscopic roughness described by two parameters, the AART model achieved an accuracy of about plus or minus 0.05 with a possible bias of plus or minus 0.03 in the spectral range 0.4-2.2 micrometers. This high accuracy holds at view zenith angles below 55-60 degrees covering the practically important range for remote sensing applications, and includes both glint and backscattering directions.

Lyapustin, A.↗

NASA NTRSDate not supplied

Statistical Classification of Biosignature Information using Multiple Instrument Observations

The accurate identification of biosignatures (indications of life) from data taken from remote or in situ planetary exploration is one of the most important challenges in astrobiology, the interdisciplinary field examining habitability and the potential for extraterrestrial life. This study employs machine learning algorithms to optimize the identification of biosignatures, with an emphasis on those which are agnostic to a specific biochemical basis. We exploit the wealth of terrestrial data available from biogenic and abiogenic systems to enhance efficient feature prioritization. Our dataset, pulled from public databases and laboratory recorded measurements, includes elemental abundance, isotopic fractionation, and VNIR/Raman spectra The data curation process included standardization for detection limits and ranges. Subsequent feature extraction yielded detailed inputs for machine learning, including combinations of elemental content, isotopic ratios, and parameters of spectral peaks and troughs. Feature significance was evaluated across diverse machine learning methodologies, such as k-nearest neighbors, logistic regression, Random Forest, support vector machines, and Gaussian Naïve Bayes, along with a combined voting classifier. We utilized Receiver Operating Characteristic Area Under the Curve (ROC AUC) across 2,000 50% test-train splits as a robust metric of model performance. Results revealed a promising ROC AUC of 0.853 for the combined voting classifier. Removing elemental abundance data notably reduced model accuracy (13% decrease in AUC), highlighting its critical role in biosignature detection. Several other individual data features exhibited significance within their respective data types, offering additional granularity. This research fortifies the relevance of machine learning to astrobiology, potentially enhancing life detection missions by allowing algorithmic prioritization of high-interest samples for further investigation. Future work will refine data standardization, expand the dataset to include more terrestrial systems, and incorporate convolutional neural networks for spectral feature extraction. The potential for public data sharing is also under exploration, reinforcing our commitment to collective scientific advancement.

Statistical↗

NASA NTRS2021

Scheduling the NASA Deep Space Network with Deep Reinforcement Learning

With three complexes spread evenly across the Earth, NASA’s Deep Space Network (DSN) is the primary means of communications as well as a significant scientific instrument for dozens of active missions around the world. A rapidly rising number of spacecraft and increasingly complex scientific instruments with higher bandwidth requirements have resulted in demand that exceeds the network’s capacity across its 12 antennae. The existing DSN scheduling process operates on a rolling weekly basis and is time-consuming; for a given week, generation of the final baseline schedule of spacecraft tracking passes takes roughly 5 months from the initial requirements submission deadline, with several weeks of peer-to-peer negotiations in between. This paper proposes a deep reinforcement learning (RL) approach to generate candidate DSN schedules from mission requests and spacecraft ephemeris data with demonstrated capability to address real-world operational constraints. A deep RL agent is developed that takes mission requests for a given week as input, and interacts with a DSN scheduling environment to allocate tracks such that its reward signal is maximized. A comparison is made between an agent trained using Proximal Policy Optimization and its random, untrained counterpart. The results represent a proof-of-concept that, given a well-shaped reward signal, a deep RL agent can learn the complex heuristics used by experts to schedule the DSN. A trained agent can potentially be used to generate candidate schedules to bootstrap the scheduling process and thus reduce the turnaround cycle for DSN scheduling.

Wilson, Brian↗