Optimizing Deep Learning Models for On-Orbit Deployment Through Neural Architecture Search

Optimizing Deep Learning Models for On-Orbit Deployment Through Neural Architecture Search | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Optimizing Deep Learning Models for On-Orbit Deployment Through Neural Architecture Search Roberto Del Prete, Parampuneet Kaur Thind, Andrea Mazzeo, Matthew Whitley, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6753230/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 29 Oct, 2025 Read the published version in Scientific Reports → Version 1 posted 11 You are reading this latest preprint version Abstract Advancements in spaceborne edge computing has facilitated the incorporation of Artificial Intelligence (AI)-powered chips into CubeSats, allowing for intelligent data handling and enhanced analytical capabilities with greater operational autonomy. This class of satellites face stringent energy and memory constraints, thus necessitating lightweight models which are often obtained by compression techniques. This paper addresses model compression by Neural Architecture Search (NAS) to enable computational efficiency and balance between accuracy, size, and latency. More in detail, we design an evolutionary-based NAS framework for onbord processing and test its capabilities on the burned area segmentation test case. The proposed solution jointly optimizes network architecture and deployment for hardware-specific resource-constrained platforms. Additionally, hardware-awareness is introduced in the optimization loop for tailoring the network topology to the specific target edge computing chip. The resulting models, which has been desiged on CubeSat-class hardware, i.e. an NVIDIA Jetson AGX Orion and the Intel Movidious Myriad X, exhibits a memory footprint below 1MB, outperforming handcrafted baselines in terms of latency (3°ø faster) and maintain competitive mean Intersection over Union (mIoU); additionally enabling real-time, high-resolution inference in orbit. Earth and environmental sciences/Natural hazards Physical sciences/Engineering Full Text Additional Declarations No competing interests reported. Supplementary Files supplementary.zip Cite Share Download PDF Status: Published Journal Publication published 29 Oct, 2025 Read the published version in Scientific Reports → Version 1 posted Editorial decision: Revision requested 07 Aug, 2025 Reviews received at journal 05 Aug, 2025 Reviewers agreed at journal 02 Aug, 2025 Reviewers agreed at journal 31 Jul, 2025 Reviews received at journal 16 Jul, 2025 Reviewers agreed at journal 14 Jul, 2025 Reviewers invited by journal 02 Jul, 2025 Editor assigned by journal 02 Jul, 2025 Editor invited by journal 03 Jun, 2025 Submission checks completed at journal 30 May, 2025 First submitted to journal 30 May, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6753230","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":471122049,"identity":"37259917-03c8-479a-a039-06ffe01c459c","order_by":0,"name":"Roberto Del Prete","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABAUlEQVRIiWNgGAWjYFACHih9gLkBTPODSQNCWhJAWhgbwEolGyBaJIjXYnAAIo5Ti3l778HHlT9sGPiON7Y9rqj4I2d8vP3hB4aCO3W4tMicOZdseCYhjUHyzMF2wzNnDIzNzpwxlmAweIbTFgmJHDPJhoTDDAY3EtskG9sMErfdyAE6yuAwPi3mP8Fa7j8EavlnUL95/vPHPwhoMWOE2MII1NJgkGAgwWCG3xaec8mSDWlpPJJnEtsNG44ZG844k2NmkWBwGBLY2LSw9x782GBjI8d3/PCxhw01cvL87ccf3/jw5zA/LltgABQ7bAhuAiH1UMBGWMkoGAWjYBSMSAAAZwhX55I4JLAAAAAASUVORK5CYII=","orcid":"","institution":"European Space Research Institute","correspondingAuthor":true,"prefix":"","firstName":"Roberto","middleName":"Del","lastName":"Prete","suffix":""},{"id":471122050,"identity":"bc3f8842-0618-40ce-a9bf-11b0fda49c5a","order_by":1,"name":"Parampuneet Kaur Thind","email":"","orcid":"","institution":"Little Place Labs","correspondingAuthor":false,"prefix":"","firstName":"Parampuneet","middleName":"Kaur","lastName":"Thind","suffix":""},{"id":471122053,"identity":"3d7141b8-7346-4712-819d-0b0cb47cadb7","order_by":2,"name":"Andrea Mazzeo","email":"","orcid":"","institution":"University of Naples Federico II","correspondingAuthor":false,"prefix":"","firstName":"Andrea","middleName":"","lastName":"Mazzeo","suffix":""},{"id":471122054,"identity":"6268920c-cbf7-4052-a687-c7b5cc152764","order_by":3,"name":"Matthew Whitley","email":"","orcid":"","institution":"Little Place Labs","correspondingAuthor":false,"prefix":"","firstName":"Matthew","middleName":"","lastName":"Whitley","suffix":""},{"id":471122057,"identity":"9f95fe27-e602-48de-b920-9efdaf44e028","order_by":4,"name":"Lorenzo Papa","email":"","orcid":"","institution":"European Space Research Institute","correspondingAuthor":false,"prefix":"","firstName":"Lorenzo","middleName":"","lastName":"Papa","suffix":""},{"id":471122058,"identity":"8f794619-d1a5-4138-9aac-8fdfc79f9e71","order_by":5,"name":"Nicolas Longepe","email":"","orcid":"","institution":"European Space Research Institute","correspondingAuthor":false,"prefix":"","firstName":"Nicolas","middleName":"","lastName":"Longepe","suffix":""},{"id":471122059,"identity":"69de8db5-fba5-40cb-a115-01633f67ae46","order_by":6,"name":"Gabriele Meoni","email":"","orcid":"","institution":"European Space Research Institute","correspondingAuthor":false,"prefix":"","firstName":"Gabriele","middleName":"","lastName":"Meoni","suffix":""}],"badges":[],"createdAt":"2025-05-26 18:38:12","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6753230/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6753230/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41598-025-21467-8","type":"published","date":"2025-10-29T15:58:05+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":95040016,"identity":"99f35214-9508-4241-9a7a-a0f39b406fe9","added_by":"auto","created_at":"2025-11-03 16:07:27","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":6267218,"visible":true,"origin":"","legend":"","description":"","filename":"Article.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6753230/v1_covered_20141b5f-d6b3-4707-8a54-3d31e4cda9a6.pdf"},{"id":84664975,"identity":"1e56d61c-c502-4347-8492-def37c8fdc04","added_by":"auto","created_at":"2025-06-16 05:33:33","extension":"zip","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":1848589,"visible":true,"origin":"","legend":"","description":"","filename":"supplementary.zip","url":"https://assets-eu.researchsquare.com/files/rs-6753230/v1/aa208215a5345251766af8bf.zip"}],"financialInterests":"No competing interests reported.","formattedTitle":"Optimizing Deep Learning Models for On-Orbit Deployment Through Neural Architecture Search","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-6753230/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6753230/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Advancements in spaceborne edge computing has facilitated the incorporation of Artificial Intelligence (AI)-powered chips into CubeSats, allowing for intelligent data handling and enhanced analytical capabilities with greater operational autonomy. This class of satellites face stringent energy and memory constraints, thus necessitating lightweight models which are often obtained by compression techniques. This paper addresses model compression by Neural Architecture Search (NAS) to enable computational efficiency and balance between accuracy, size, and latency. More in detail, we design an evolutionary-based NAS framework for onbord processing and test its capabilities on the burned area segmentation test case. The proposed solution jointly optimizes network architecture and deployment for hardware-specific resource-constrained platforms. Additionally, hardware-awareness is introduced in the optimization loop for tailoring the network topology to the specific target edge computing chip. The resulting models, which has been desiged on CubeSat-class hardware, i.e. an NVIDIA Jetson AGX Orion and the Intel Movidious Myriad X, exhibits a memory footprint below 1MB, outperforming handcrafted baselines in terms of latency (3°ø faster) and maintain competitive mean Intersection over Union (mIoU); additionally enabling real-time, high-resolution inference in orbit.","manuscriptTitle":"Optimizing Deep Learning Models for On-Orbit Deployment Through Neural Architecture Search","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-06-16 05:33:28","doi":"10.21203/rs.3.rs-6753230/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-08-07T06:50:22+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-08-05T17:40:59+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"326151174290509037326861827493852849421","date":"2025-08-02T07:15:07+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"222015564362706607173159857788187427965","date":"2025-07-31T09:26:34+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-07-16T13:08:58+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"24937928946857733614730731409736001403","date":"2025-07-14T10:42:01+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-07-02T09:19:58+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-07-02T09:17:52+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-06-03T13:31:20+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-05-30T07:09:14+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2025-05-30T07:06:01+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"ac6102fb-d33e-4d19-b4ba-810e9af1b39d","owner":[],"postedDate":"June 16th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":50035847,"name":"Earth and environmental sciences/Natural hazards"},{"id":50035848,"name":"Physical sciences/Engineering"}],"tags":[],"updatedAt":"2025-11-03T16:02:01+00:00","versionOfRecord":{"articleIdentity":"rs-6753230","link":"https://doi.org/10.1038/s41598-025-21467-8","journal":{"identity":"scientific-reports","isVorOnly":false,"title":"Scientific Reports"},"publishedOn":"2025-10-29 15:58:05","publishedOnDateReadable":"October 29th, 2025"},"versionCreatedAt":"2025-06-16 05:33:28","video":"","vorDoi":"10.1038/s41598-025-21467-8","vorDoiUrl":"https://doi.org/10.1038/s41598-025-21467-8","workflowStages":[]},"version":"v1","identity":"rs-6753230","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6753230","identity":"rs-6753230","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}