[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-118472-en":3,"doc-seo-118472-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},118472,1374391974468,"Eden","https://ap-avatar.wpscdn.com/davatar_29158cc5080c5b710cf443261637dec0",6,"Technology","Unveiling the Potential of Tiny Machine Learning for Enhanced People Counting in UWB Radar Data","Tiny Machine Learning (TinyML) shifts intelligence processing close to data generation, lowering decision latency and enabling inference even without reliable remote connectivity. In this setting, Ultra-Wideband (UWB) radar data provide a challenging yet privacy-preserving sensing source. The work proposes a TinyML method that counts people in a given area by processing UWB radar measurements. The solution targets high counting accuracy while reducing memory and computation so it can run on tiny edge devices, with experimental validation on a real-world dataset.","Unveiling the Potential of Tiny Machine Learning for Enhanced People Counting in UWB Radar Data  \nMassimo Pavan 1 , Luis Gonz´alez Navarro3 , Armando Caltabiano2 , and Manuel  \nRoveri 1  \n1 Politecnico di Milano, Milano, IT {massimo.pavan,[manuel.roveri](manuel.roveri}@polimi.it)[}](manuel.roveri}@polimi.it)[@polimi.it](manuel.roveri}@polimi.it)[ ](manuel.roveri}@polimi.it)2 Truesense s.r.l. , Milano, [IT](IT armando.caltabiano@truesense.it)[ armando.caltabiano@truesense.it](IT armando.caltabiano@truesense.it)  \n3 Universidad Polit´ecnica de Madrid, Madrid, ES [luis.gnavarro@alumnos.upm.es](luis.gnavarro@alumnos.upm.es)  \nAbstract. Tiny Machine Learning (TinyML) allows to move the intelligence processing as close as possible to where data are generated, hence reducing the latency with which a decision is made and being able to process data even when remote connection is scarce or absent. In this technological scenario, Ultra-Wideband (UWB) radar data represent anew and challenging source of data providing relevant information, while guaranteeing the privacy of users. This paper introduces a novel TinyML solution able to count the number of people in a given area by processing UWB radar data. This novel solution was carefully designed to guarantee a high counting accuracy, while reducing the memory and computational demand so as to be executed on tiny devices. Experimental results on a real-world UWB radar dataset show the effectiveness of the proposed solution.  \nKeywords: Tiny Machine Learning · Ultra-Wideband (UWB) radar · People Counting.  \n1 Introduction  \nIn the most recent technological landscape, tiny devices are becoming one of the main areas of technological breakthrough. As Internet-of-Things (IoT) units, embedded systems, and edge devices become more present in the technological environment, the scientific trend reflects the displacement of data processing closer to where the data are generated. This aims to increase the autonomy of tiny devices since decisions can be taken locally, while reducing the energy consumption (since transmitting is far more energy-demanding than processing the data), the required bandwidth, and the latency of decision-making. Designing Machine Learning (ML) models meant to operate on tiny devices requires to completely redesign the traditional ML models and algorithms due to the severe technological constraints on memory, computation and energy consumption [27, 4,5] .  \n2 M. Pavan, L. Navarro et al.  \nThis is exacly where Tiny Machine Learning (TinyML) comes into play by introducing tiny models and algorithms as well as approximate computing mechanisms (e.g., quantization and pruning) to design ML solutions able to satisfy the aforementioned technological constraints of tiny devices [23,2] .  \nOne of the most promising applications for TinyML is presence detection, aiming at identifying the presence of one or more persons in a given environment. However, current solutions rely on the processing of images taken from video cameras or microphones [26,7], hence potentially impacting on the privacy of users.  \nIn order to deal with and address this privacy issue in detecting and counting people, we introduce the use of Ultra-Wideband (UWB) radar, which shows tobe a promising radar technology for human activity recognition [10,15] . From the technological point-of-view, UWB radar devices are characterized by precise recordings (in the order of the mm), low energy consumption (generally below 0.1 W) and fast acquisition of data (in fractions of seconds), making them valuable solutions even for tiny devices. In this perspective, TinyML solutions for UWBradar will pave the way for a large number of interesting applications based on tiny devices and will guarantee the privacy of users. Very few examples of TinyML solutions operating on UWB radar data are available in the literature [20,19] and currently none of these solutions addresses the problem of people counting.  \nThe aim of this paper is to ","cbCaia2rI8GEQ0pc","https://ap.wps.com/l/cbCaia2rI8GEQ0pc","pdf",443060,1,17,"English","en",105,"# Introduction\n## Motivation for TinyML on Edge Devices\n## Privacy-Preserving People Counting Motivation\n# Proposed Approach\n## Preprocessing for Radar Data\n## Lightweight Tiny Dilated CNN Design\n## Quantization and Model Compression\n# Problem Formulation and Dataset Setup\n## Regression-Based People Counting\n## Car Backseat Passenger Scenario\n# Evaluation and Results\n## Memory and Computational Demand Assessment\n## Effectiveness on Real-World UWB Data","[{\"question\":\"What problem does the paper address?\",\"answer\":\"It introduces a TinyML solution to count the number of people in an area using UWB radar data, including a scenario focused on counting passengers inside a car.\"},{\"question\":\"Why use UWB radar instead of cameras or microphones?\",\"answer\":\"UWB radar supports human activity recognition while helping preserve user privacy, whereas camera/video and microphone-based approaches may raise privacy concerns.\"},{\"question\":\"How does the proposed method fit TinyML constraints?\",\"answer\":\"It uses a preprocessing stage plus a custom lightweight CNN with tiny dilated convolutional blocks and quantization mechanisms to reduce memory and computational demand for deployment on tiny devices.\"}]","Unveiling the Potential of Tiny Machine Learning for Enhanced People Counting in UWB Radar Data | PDF",1785683774,43,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"unveiling-the-potential-of-tiny-machine-learning-for-enhanced-people-counting-in-uwb-radar-data","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/technology/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/unveiling-the-potential-of-tiny-machine-learning-for-enhanced-people-counting-in-uwb-radar-data/118472/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-02",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does the paper address?","Question",{"text":75,"@type":76},"It introduces a TinyML solution to count the number of people in an area using UWB radar data, including a scenario focused on counting passengers inside a car.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"Why use UWB radar instead of cameras or microphones?",{"text":80,"@type":76},"UWB radar supports human activity recognition while helping preserve user privacy, whereas camera/video and microphone-based approaches may raise privacy concerns.",{"name":82,"@type":73,"acceptedAnswer":83},"How does the proposed method fit TinyML constraints?",{"text":84,"@type":76},"It uses a preprocessing stage plus a custom lightweight CNN with tiny dilated convolutional blocks and quantization mechanisms to reduce memory and computational demand for deployment on tiny devices.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,113,118,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":111,"slug":112},50,"technology",{"id":114,"doc_module":4,"doc_module_name":46,"category_name":115,"show_sort_weight":116,"slug":117},7,"Healthcare",40,"healthcare",{"id":119,"doc_module":4,"doc_module_name":46,"category_name":120,"show_sort_weight":121,"slug":122},8,"Research & Report",30,"research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]