Machine Learning Applications for Space Operations
| dc.contributor.author | Lan, Wenschel | |
| dc.contributor.author | Karpenko, Mark | |
| dc.contributor.corporate | Naval Research Program (NRP) | |
| dc.contributor.corporate | Space Systems Academic Group (SSAG) | |
| dc.contributor.department | Mechanical and Aerospace Engineering (MAE) | |
| dc.date.accessioned | 2025-04-10T20:22:08Z | |
| dc.date.available | 2025-04-10T20:22:08Z | |
| dc.date.issued | 2024-04-16 | |
| dc.description | NPS NRP Executive Summary | |
| dc.description.abstract | Due to the increasingly contested radio frequency (RF) environment, new approaches for satellite communications (SATCOM) interference detection and identification are needed. One approach is to apply machine-learning (ML) techniques, which can be applied on both the ground and space segments. This study focuses on exploring applications that can be supported by high-power computing, cloud-enabled services, and embedded ML hardware at the tactical edge. This study uses digitized RF transmissions to support the development of ML concepts for RF interference detection and identification. Due to the memory requirements for RF waveforms, the Cloud environment Azure was chosen to process the digitized waveforms and train the ML interference detectors. Azure shows promise in transitioning ML models to an austere environment (i.e., a ship) for processing large data sets. In this application, it was discovered that feedforward autoencoders can be used to quickly and correctly identify RF interference applied at various signal-to-noise ratios (SNRs). The statistic of corrupted signals may be classified to determine the specific type of interface that has been applied. Other ML architectures, such as long-short-term memory (LSTM) and convolutional neural networks (CNNs) are also applicable as they can process temporal features differently than an autoencoder network. Upon detection of jamming or signal degradation, it may be necessary to steer a SATCOM system to reestablish degraded and/or broken links. This aspect can be handled by performing a rapid spacecraft reorientation maneuver. The current state of practice involves developing appropriate maneuver on the ground for uplink and execution on board the vehicle. To establish autonomy of operation, a concept for on-orbit (edge) command generation using neural networks was demonstrated that can support event-driven and autonomous operations for communication applications that benefit the warfighter. | |
| dc.description.distributionstatement | Approved for public release. Distribution is unlimited. | |
| dc.description.funder | This research is supported by funding from the Naval Postgraduate School, Naval Research Program (PE 0605853N/2098). https://nps.edu/nrp | |
| dc.description.funder | Chief of Naval Operations (CNO) | |
| dc.description.sponsorship | N2/N6 - Information Warfare | |
| dc.format.extent | 5 p. | |
| dc.identifier.other | NPS-23-F277-A | |
| dc.identifier.uri | https://hdl.handle.net/10945/73595 | |
| dc.publisher | Monterey, CA; Naval Postgraduate School | |
| dc.relation.ispartofseries | Naval Research Program (NRP) Project Documents | |
| dc.rights | This publication is a work of the U.S. Government as defined in Title 17, United States Code, Section 101. Copyright protection is not available for this work in the United States. | |
| dc.subject.author | machine learning | |
| dc.subject.author | radio frequency | |
| dc.subject.author | RF | |
| dc.subject.author | interference detection | |
| dc.subject.author | space communications | |
| dc.subject.author | on-orbit processing | |
| dc.subject.author | tactical edge processing | |
| dc.title | Machine Learning Applications for Space Operations | |
| dc.type | Report | |
| dspace.entity.type | Publication | |
| relation.isDepartmentOfPublication | fcd0a3e4-3370-4559-99ee-72c50043ad3c | |
| relation.isDepartmentOfPublication.latestForDiscovery | fcd0a3e4-3370-4559-99ee-72c50043ad3c | |
| relation.isOrgUnitOfPublication | 78e1a7f9-aa70-4b0c-8c2c-4ff04808d293 | |
| relation.isOrgUnitOfPublication | 71422806-0ca0-4772-840b-075e219b82d9 | |
| relation.isOrgUnitOfPublication.latestForDiscovery | 78e1a7f9-aa70-4b0c-8c2c-4ff04808d293 | |
| relation.isSeriesOfPublication | 2c25ebbb-dea7-48fc-8f19-5278e52aa428 | |
| relation.isSeriesOfPublication.latestForDiscovery | 2c25ebbb-dea7-48fc-8f19-5278e52aa428 |
