[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-122352-en":3,"doc-seo-122352-105":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},122352,2336464648746,"Skyler","https://ap-avatar.wpscdn.com/davatar_276721f389ce27ea32af1340a28f341c",8,"Research & Report","Improved IMU/GNSS EKF fusion using Machine Learning","IMU sensors are widely used in navigation systems, but low-cost MEMS IMUs introduce larger stochastic errors that accumulate over time and cause navigation drift. Integrating GNSS measurements with IMU improves accuracy yet degrades during GNSS outages because IMU errors persist. This thesis presents an ML-driven strategy to correct IMU errors using LightGBM and CatBoost, leveraging inverse kinematics to extract clean IMU training data from EKF-estimated PVA when GNSS is available. Tree-based denoising improves position, velocity, and attitude and is faster than CNN in training and prediction.","Improved IMU/GNSS EKF fusion using Machine Learning  \nby  \nRohan Kumar Reddy Damagatla  \nA thesis submitted to the Faculty of Graduate and Postdoctoral Affairs in partial fulfillment of the requirements for the degree of  \nMaster of Applied Science  \nin  \nElectrical and Computer Engineering  \nCarleton University  \nOttawa, Ontario  \n© 2023 Rohan Kumar Reddy Damagatla  \nAbstract  \nThe applications of Inertial Measurement Unit (IMU) sensors are extended to several domains, abundantly in navigation systems. IMU sensors are the essential components in navigation systems. Extensive research in the area of IMU sensors led to the emergence of Micro-Electro-Mechanical System (MEMS) sensors as a more affordable and lightweight IMU. However, utilization of MEMS IMUs to obtain reliable navigation data is challenging as the MEMS suffer from larger stochastic errors that accumulate over time, eventually causing drifts in navigation. The integration of Global Navigation Satellite System (GNSS) measurements with IMU data solves the issue of navigation drift and provides accurate navigation. However, this integration falls short in delivering continuous and dependable navigation during GNSS outage scenarios, primarily due to persistent IMU errors. This thesis introduces an effective strategy to reduce navigation drifts by rectifying IMU errors using Machine Learning (ML) algorithms such as Light Gradient Boosting Machine (LightGBM) and Categorical Boosting (CatBoost) . In contrast to many other methodologies reliant on high-end and expensive IMUs for denoising low-cost MEMS IMUs, this thesis proposes Inverse Kinematics (IK) . IK is employed to extract pristine IMU training data from Position, Velocity, Attitude (PVA) values estimated through the Extended Kalman Filter (EKF) when GNSS signals are accessible and reliable. The distinctive advantage of the IK  \napproach lies in its capacity to obtain real-time clean IMU data without necessitating high-end IMUs for training ML models. The proposed method undergoes testing in Loosely coupled and Tightly coupled schemes using simulation and real data sets under varying GNSS outage durations. The denoised IMU signals are compared with that of conventional signal processing techniques of Moving Average (MA) and Savitzky Golay (SG) . In addition, a comparison of results between the proposed algorithms against Convolutional Neural Networks (CNN) algorithm is presented in this thesis. The outcomes showcase a substantial enhancement in position, velocity, and orientation estimation. The computation time required for the model training and prediction of various algorithms is compared and the results show the superiority of the proposed tree-based algorithms over conventional CNN architecture.  \nAcknowledgements  \nFirstly, I want to extend my gratitude and thanks to my supervisor, Prof. Mohamed Atia; without him, this thesis would not have been possible. His valuable input and discussions helped me successfully complete my thesis. His patience and belief in me are significant contributing factors that helped me at various stages of my thesis.  \nSpecial thanks to my fellow researchers Sarat, Hari and Jonathan from Embedded and Multi-sensor Systems Lab for showing interest in my thesis and having stipulated conversations which helped me in progressing the thesis.  \nI am always grateful for my parents and family, who have supported me in fulfilling my dreams. I would also like to thank my friends Anurag, Vishnu, Aswini, Rekha and Aarthi for all their support and constant motivation throughout my Masters duration and for making this journey memorable.  \nContents  \nAbstract i  \nAcknowledgements iii  \nList of Tables vii  \nList of Figures ix  \nAcronyms xii  \n1 Introduction 1  \n1.1 Objective ................................. 2  \n1.2 Proposed Methodology .......................... 3  \n1.3 Contributions ............................... 4  \n1.4 List of Publications ............................ 5  \n1.5 Thesis Structure .","cbCaiiuFKnxPCpEk","https://ap.wps.com/l/cbCaiiuFKnxPCpEk","pdf",3322110,1,114,"English","en",105,"# Abstract\n# Introduction\n## Objective\n## Proposed Methodology\n## Contributions\n## Thesis Structure\n# Background and Related Works\n## IMU Sensors\n## Global Navigation Satellite System\n## Navigation Systems\n## IMU and GNSS Fusion\n## Related Works\n# Denoising IMU using Inverse Kinematics and Machine Learning\n## IMU Mechanization Equations\n## Inverse Kinematics\n## Gradient Boosting Algorithms\n## Proposed Methodology\n# ML-based IMU denoising for Loosely Coupled Integration\n## IMU GNSS Loosely Coupled Fusion\n## Experiment Setup\n## Results and Analysis","[{\"question\":\"Why does IMU/GNSS fusion suffer during GNSS outage scenarios?\",\"answer\":\"GNSS outage prevents reliable satellite measurements, so the system relies on IMU data. Persistent IMU errors accumulate and cause drift, reducing navigation continuity and dependability.\"},{\"question\":\"How does inverse kinematics help obtain training data for machine learning?\",\"answer\":\"Inverse kinematics extracts cleaner IMU training signals by using PVA values estimated from the Extended Kalman Filter when GNSS is available and reliable.\"},{\"question\":\"Which denoising approaches are compared in the thesis, and what is the key result?\",\"answer\":\"The proposed LightGBM and CatBoost tree-based algorithms are compared against Moving Average and Savitzky-Golay filtering, and against a Convolutional Neural Network. Results show substantial improvement in position, velocity, and orientation, with better computation time than the CNN approach.\"}]","Improved IMU/GNSS EKF fusion using Machine Learning | PDF",1785810189,287,{"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},"improved-imugnss-ekf-fusion-using-machine-learning","",{"@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/improved-imugnss-ekf-fusion-using-machine-learning/122352/",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","2026-08-04",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},"Why does IMU/GNSS fusion suffer during GNSS outage scenarios?","Question",{"text":75,"@type":76},"GNSS outage prevents reliable satellite measurements, so the system relies on IMU data. Persistent IMU errors accumulate and cause drift, reducing navigation continuity and dependability.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does inverse kinematics help obtain training data for machine learning?",{"text":80,"@type":76},"Inverse kinematics extracts cleaner IMU training signals by using PVA values estimated from the Extended Kalman Filter when GNSS is available and reliable.",{"name":82,"@type":73,"acceptedAnswer":83},"Which denoising approaches are compared in the thesis, and what is the key result?",{"text":84,"@type":76},"The proposed LightGBM and CatBoost tree-based algorithms are compared against Moving Average and Savitzky-Golay filtering, and against a Convolutional Neural Network. Results show substantial improvement in position, velocity, and orientation, with better computation time than the CNN approach.","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},[93,97,101,105,110,115,120,123,128,131,135],{"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":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":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"]