[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-118419-en":3,"doc-seo-118419-105":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},118419,1099514067438,"River Wang","https://ap-avatar.wpscdn.com/avatar/100002539ee87300030?x-image-process=image/resize,m_fixed,w_180,h_180&k=1780474512215547542",8,"Research & Report","Efficient machine learning for motion sensing for lighting applications - Research paper","Machine learning-based signal classification for motion sensing in lighting systems faces two key difficulties: strongly skewed datasets and stringent optimization requirements tied to response time and false-positive behavior. This paper presents an efficient, fully automated method using a customized loss function and a product of probability models, yielding probability models with reduced complexity. The resulting approach supports accurate, resource-friendly classification suitable for embedded motion-sensing applications.","arXiv :2406 . 16723v1 [ ee ss . SP] 21 Jun 2024  \nEfficient machine learning for motion sensing for lighting applications  \nFetze Pijlman  \nSignify Research, Technical University Eindhoven  \nJune 25, 2024  \nAbstract  \nThe use of machine learning for building a classifier in signal processing for motion sensing presents unique challenges. This paper proposes a novel method that effectively addresses the combination of skewed datasets and optimization requirements. By utilizing a customized loss function and a product of probability models, our approach achieves a fully automated and efficient machine learning process. Additionally, our resulting probability models offer reduced complexity, making them ideal for embedded applications. Our method offers a promising solution for motion sensing applications that require accurate and efficient classification.  \n1 Introduction  \n1.1 Motion sensing for lighting applications  \nMotion sensing is a popular method for controlling lights in the environment. A popular control scheme contains a first ingredient being that when motion is observed while the lamp being off, the lamp will be switched on for a predefined length of time (also known as hold time which is tracked by a counter) . Typical values for hold time are in the range of 5 to 20 minutes. A second ingredient is that when motion is being observed while the lamp being on, the counter that tracks the amount of time that has passed is being reset. The requirements for the first ingredient are demanding. For a person approaching the response time should be quick and false positives must be avoided. A typical response time is 0.6 seconds and a typical false positive rate is 1 per 1000 hours. A low false positive rate prevents a large office building to appear as a disco to an outsider late in the evening. The requirement for the second ingredient is often substantially different. The motion that needs to be sensed is often smaller in signal (turning a page in a book often leads to a smaller signal than a person walking in), response time can be slower as a reset of the counter cannot be directly observed, and the false positive rate can be higher since false positives only lead to an elongation of the time that the light is on. The detectability of a nearby arm motion should typically be above 25%[1](one is allowed to miss  \nFigure 1: Some signals related to motion sensing. The left and center graph are typical motion and noise signals, the right graph is an example of a spurious noise signal.  \nsome motions without leading to drastic consequences) and the false positive rate should typically be below 1 per 20 hours.  \nWhen developing motion sensors one usually records a set of motion events and a substantial amount of hours of so-called noise. The motion events are typically recorded at various angles and distances from the motion sensor. A data set containing a few thousands of these recorded events is very typical for developing a motion sensor. The collection of noise data is different. Where the collection of motion data can be done by following a protocol, the collection of noise data is often done by long measurements. The signals that are collected in these measurements are most often small in size but the appearance of occasional spurious noise signals (e.g. due to radio interference) is common. Some example signals are presented in Fig. 1. The recording and handling of occasional spurious noise signals makes the development of a signal processing algorithm challenging.  \n1.2 Challenges for using machine learning  \nA typical signal processing algorithm is fed by a time series of regularly sampled samples on which the algorithm needs to respond. Algorithms such as recurrent neural networks and Bayesian models are of interest although being of the infinite response type (classification depends on full history) . Another option is to group samples and to classify the group, an approach that has simularities with FIR filters. Th","cbCainaY4DBMcZVZ","https://ap.wps.com/l/cbCainaY4DBMcZVZ","pdf",396629,1,9,"English","en",105,"# Introduction\n## Motion sensing for lighting applications\n## Challenges for using machine learning","[{\"question\":\"What makes machine learning for lighting motion sensing especially challenging?\",\"answer\":\"The task combines strongly skewed motion/noise datasets with strict optimization requirements for response time and false-positive rates.\"},{\"question\":\"What approach does the paper propose to handle skewed data and optimization needs?\",\"answer\":\"It uses a customized loss function and a product of probability models to build a fully automated and efficient learning process.\"},{\"question\":\"Why are the resulting probability models useful for real lighting hardware?\",\"answer\":\"The probability models are designed to have reduced complexity, making them suitable for embedded applications where computation and size are limited.\"}]","Efficient machine learning for motion sensing for lighting applications - Research paper | PDF",1785683523,23,{"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},"efficient-machine-learning-for-motion-sensing-for-lighting-applications-research-paper","",{"@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/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/efficient-machine-learning-for-motion-sensing-for-lighting-applications-research-paper/118419/",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 makes machine learning for lighting motion sensing especially challenging?","Question",{"text":75,"@type":76},"The task combines strongly skewed motion/noise datasets with strict optimization requirements for response time and false-positive rates.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"What approach does the paper propose to handle skewed data and optimization needs?",{"text":80,"@type":76},"It uses a customized loss function and a product of probability models to build a fully automated and efficient learning process.",{"name":82,"@type":73,"acceptedAnswer":83},"Why are the resulting probability models useful for real lighting hardware?",{"text":84,"@type":76},"The probability models are designed to have reduced complexity, making them suitable for embedded applications where computation and size are limited.","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},[93,97,101,105,110,115,120,123,127,130,134],{"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":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":21,"doc_module":4,"doc_module_name":46,"category_name":124,"show_sort_weight":125,"slug":126},"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]