This dissertation reports on the design and implementation of a high-frequency, tri-axial capacitive resonant gyroscopes integrated on a single chip. The components that construct tri-axial rotation sensing consist of a yaw, a pitch and a roll device. The yaw-rate gyroscope has a wide bandwidth and a large full-scale range, and operates at a mode-matched condition with DC polarization voltage of 10V without frequency tuning requirement. The large bandwidth of 3kHz and expected full-scale range over 30,000˚/sec make the device exhibit fast rate response for rapid motion sensing application. For the pitch-and-roll rate sensing, an in-plane drive-mode and two orthogonal out-of-plane sense-modes are employed. The rotation-rate sensing from lateral axes is performed by mode-matching the in-plane drive-mode with out-of-plane sense-modes to detect Coriolis-force induced deflection of the resonant mass. To compensate process variations and thickness deviations in the employed silicon-on-insulator (SOI) substrates, large electrostatic frequency tunings of both the drive and sense modes are realized. A revised high aspect ratio combined polysilicon and silicon (HARPSS) process is developed to resolve the Coriolis response that exists toward out-of-plane direction while drive-mode exists on in-plane, and tune individual frequencies with minimal interference to unintended modes. To conclude and overcome the performance limitation, design optimization of high-frequency tri-axial gyroscopes is suggested. Q-factor enhancement through reduction of thermoelastic damping (TED) and optimizations of physical dimensions are suggested for the yaw disk gyroscope. For the pitch-and-roll gyroscope, scaling property of physical dimension and its subsequent performance enhancement are analyzed.

L'estimation de l'attitude (ou orientation) est un problème récurrent de nombreuses applications allant de la robotique aérienne ou sous-marine en passant par des applications médicales (surveillance de patients, réhabilitation), mais aussi jeux vidéo, etc. L'objectif de cette thèse est d'évaluer l'apport des approches ensemblistes dans le cadre de l'estimation de l'attitude à partir de données issues de triaxes accéléromètres (A), magnétomètres (M) et gyromètres (G). Dans un premier temps, on s'intéresse aux mouvements "quasi-statiques" et l'estimation de l'attitude est réalisée à partir de mesures AM. On aborde ensuite le cas des mouvements dynamiques, en considérant l'ensemble des mesures AGM. Le problème du choix de la paramétrisation de l'attitude a été abordé et on a comparé les résultats obtenus et le temps calcul pour des modélisations avec les angles de Cardan et le quaternion unitaire. Les algorithmes développés ont été validés en simulation et avec des données réelles. Les résultats ont été comparés avec ceux fournis par des algorithmes de l'état de l'art, par exemple SIVIA. La deuxième partie du manuscrit est consacrée à au diagnostic des capteurs de la centrale inertielle avec des approches ensemblistes. Les algorithmes développés dans la première partie du travail sont adaptés afin de pouvoir détecter et localiser un défaut dans l'ensemble des capteurs considérés. / Attitude estimation is one of the prominent problem encountered in various application areas such as Aerial and submarine robotics, bio-medical applications (elderly people monitoring, rehabilitation) but also, video game and augmented reality. The main objective of this PhD is to assess the capabilities of set-membership estimation in the field of attitude estimation when triaxes accelerometer (A) magnetometer (M) and rate gyros (G) are used. Quasi-static movements are first considered. In this case AM measurements are taken into account. Then the dynamic case is considered with AGM measurement taken into account in the set-membership estimation algorithm. The problem of attitude parametrisation is also studied as it will have a strong in uence on the computational time. The algorithms proposed during this work have been validated with simulated and real data. The second part of the report deals with Fault Detection and Isolation based upon set-membership approaches. The algorithms that have been developed in the first part of this work have been adapted to cope with diagnosis of a faulty sensor within the Inertial Measurement Unit.

Este trabalho apresenta o estudo e a implementação de um sistema de navegação em tempo-real utilizado para estimar a posição, a velocidade e a atitude de um veículo submarino autônomo. O algoritmo investigado é o do Filtro de Kalman Estendido. Este filtro é freqüentemente usado para realizar a fusão de dados obtidos de diferentes sensores, em uma estimativa estatisticamente ótima, quando se respeita algumas condições. Neste trabalho, fez se a fusão entre os seguintes sensores: unidade de navegação inercial do tipo strapdown, sensor acústico de posicionamento, profundímetro, sensor de velocidade de efeito Doppler e uma bússola. Para a aplicação embarcada do Filtro de Kalman, faz-se necessário o seu desenvolvimento em tempo real. Conseqüentemente, este trabalho apresenta o estudo das principais características de um sistema de tempo real. Para desenvolver o código em C utilizou-se de algumas funções do Matlab com a finalidade de se tentar minimizar os erros de implementação do filtro. Além disto, para facilitar a implementação e respeitar os critérios de sistemas de tempo real utilizou-se de um sistema operacional, C/OS-II que possibilita aplicar sistemas com multiprocessos e utilizar semáforos para o gerenciamento do EKF, além disto, foram utilizadas normas de programação, MISRAC, para padronizar o código e aumentar a sua confiabilidade. São apresentadas também a modelagem cinemática, a metodologia e as ferramentas computacionais utilizadas para o filtro. Com base nas simulações e nos ensaios de campo executados on-line, observou-se que os filtros projetados para se estimar a atitude e a posição do veículo obtiveram bons desempenhos, além disto, foi possível verificar a convergência dos EKFs. Para estas simulações e ensaios, foram também estudados casos de situações adversas como, por exemplo, uma falha no sensor de referência de posição, sendo que para esta situação, o EKF de posição e velocidade obteve resultados satisfatórios. / This paper presents the study and implementation of a real-time navigation system used to estimate the position, velocity and attitude of an autonomous underwater vehicle. The Extended Kalman Filter, EKF, was adopted. This filter is often used to perform the data fusion from different sensors, in generating a statistically optimal estimate when some required conditions are fulfilled. The algorithm implements the fusion of the following sensors: an inertial navigation unit sensor (strapdown type), an acoustic positioning, a depth gauge, a Doppler velocity log sensor and a magnetic compass. This work presents the kinematic modelling, the methodology and computational tools used for developing the EKF algorithm. In order to integrate the EKF into an embedded system, it is necessary to develop it in real time. It was adopted the C / OS-II operational system, which allows to implement multithreaded systems and use traffic lights to manage the EKF. Furthermore, programming standards, such as MISRA C, was chosen to standardize the code and increase its reliability. The C code implementation took advantage of some Matlab functions to minimize implementation errors. Based on simulations and field tests carried out online, it was concluded that the filters designed to estimate the attitude and position of the vehicle provided good performances, in addition, it was possible to verify the EKFs convergence. The filters were tested in same adverse situations, e.g., a fault in the position reference sensor, providing satisfactory results as well.

Este trabalho apresenta o estudo e a implementação de um sistema de navegação em tempo-real utilizado para estimar a posição, a velocidade e a atitude de um veículo submarino autônomo. O algoritmo investigado é o do Filtro de Kalman Estendido. Este filtro é freqüentemente usado para realizar a fusão de dados obtidos de diferentes sensores, em uma estimativa estatisticamente ótima, quando se respeita algumas condições. Neste trabalho, fez se a fusão entre os seguintes sensores: unidade de navegação inercial do tipo strapdown, sensor acústico de posicionamento, profundímetro, sensor de velocidade de efeito Doppler e uma bússola. Para a aplicação embarcada do Filtro de Kalman, faz-se necessário o seu desenvolvimento em tempo real. Conseqüentemente, este trabalho apresenta o estudo das principais características de um sistema de tempo real. Para desenvolver o código em C utilizou-se de algumas funções do Matlab com a finalidade de se tentar minimizar os erros de implementação do filtro. Além disto, para facilitar a implementação e respeitar os critérios de sistemas de tempo real utilizou-se de um sistema operacional, C/OS-II que possibilita aplicar sistemas com multiprocessos e utilizar semáforos para o gerenciamento do EKF, além disto, foram utilizadas normas de programação, MISRAC, para padronizar o código e aumentar a sua confiabilidade. São apresentadas também a modelagem cinemática, a metodologia e as ferramentas computacionais utilizadas para o filtro. Com base nas simulações e nos ensaios de campo executados on-line, observou-se que os filtros projetados para se estimar a atitude e a posição do veículo obtiveram bons desempenhos, além disto, foi possível verificar a convergência dos EKFs. Para estas simulações e ensaios, foram também estudados casos de situações adversas como, por exemplo, uma falha no sensor de referência de posição, sendo que para esta situação, o EKF de posição e velocidade obteve resultados satisfatórios. / This paper presents the study and implementation of a real-time navigation system used to estimate the position, velocity and attitude of an autonomous underwater vehicle. The Extended Kalman Filter, EKF, was adopted. This filter is often used to perform the data fusion from different sensors, in generating a statistically optimal estimate when some required conditions are fulfilled. The algorithm implements the fusion of the following sensors: an inertial navigation unit sensor (strapdown type), an acoustic positioning, a depth gauge, a Doppler velocity log sensor and a magnetic compass. This work presents the kinematic modelling, the methodology and computational tools used for developing the EKF algorithm. In order to integrate the EKF into an embedded system, it is necessary to develop it in real time. It was adopted the C / OS-II operational system, which allows to implement multithreaded systems and use traffic lights to manage the EKF. Furthermore, programming standards, such as MISRA C, was chosen to standardize the code and increase its reliability. The C code implementation took advantage of some Matlab functions to minimize implementation errors. Based on simulations and field tests carried out online, it was concluded that the filters designed to estimate the attitude and position of the vehicle provided good performances, in addition, it was possible to verify the EKFs convergence. The filters were tested in same adverse situations, e.g., a fault in the position reference sensor, providing satisfactory results as well.

This dissertation presents a novel approach to develop a system for real-time tracking of the position and orientation of the human hand in three-dimensional space, using MEMS inertial measurement units (IMUs) and infrared cameras. This research focuses on the study and implementation of an algorithm to correct the gyroscope drift, which is a major problem in orientation tracking using commercial-grade IMUs. An algorithm to improve the orientation estimation is proposed. It consists of: 1.) Prediction of the bias offset error while the sensor is static, 2.) Estimation of a quaternion orientation from the unbiased angular velocity, 3.) Correction of the orientation quaternion utilizing the gravity vector and the magnetic North vector, and 4.) Adaptive quaternion interpolation, which determines the final quaternion estimate based upon the current conditions of the sensor. The results verified that the implementation of the orientation correction algorithm using the gravity vector and the magnetic North vector is able to reduce the amount of drift in orientation tracking and is compatible with position tracking using infrared cameras for real-time human hand motion tracking. Thirty human subjects participated in an experiment to validate the performance of the hand motion tracking system. The statistical analysis shows that the error of position tracking is, on average, 1.7 cm in the x-axis, 1.0 cm in the y-axis, and 3.5 cm in the z-axis. The Kruskal-Wallis tests show that the orientation correction algorithm using gravity vector and magnetic North vector can significantly reduce the errors in orientation tracking in comparison to fixed offset compensation. Statistical analyses show that the orientation correction algorithm using gravity vector and magnetic North vector and the on-board Kalman-based orientation filtering produced orientation errors that were not significantly different in the Euler angles, Phi, Theta and Psi, with the p-values of 0.632, 0.262 and 0.728, respectively. The proposed orientation correction algorithm represents a contribution to the emerging approaches to obtain reliable orientation estimates from MEMS IMUs. The development of a hand motion tracking system using IMUs and infrared cameras in this dissertation enables future improvements in natural human-computer interactions within a 3D virtual environment.

Global Navigation Satellite System/Inertial Measurement Unit (GNSS/IMU) Integrated Navigation Systems (INS) integrate the positive features of GNSS and IMU for optimal navigation guidance in high accuracy outdoor navigation systems, for example using Extended Kalman Filter (EKF) techniques. Time synchronization of IMU data with precise GNSS based time is necessary to accurately synchronize the two systems. This must be done in real-time for time sensitive navigation applications such as autonomous vehicles. The research is done in two parts. The first part is the simulation of inaccurate time-stamping in a single axis of nonlinear input data in a gyroscope and an accelerometer, to obtain the timing error value that is tolerable by a high accuracy GNSS/INS system. The second part is the creation of a real-time algorithm using an STM32 embedded system enabled with FreeRTOS real-time kernel for a GNSS receiver and antenna, along with an IMU sensor. A comparative analysis of the time synchronized system and an unsynchronized system is done based on the errors produced using gyroscope and accelerometer readings along a single axis from the IMU sensor, by conducting static and rotational tests on a revolving chair.The simulation concludes that a high accuracy GNSS/INS system can tolerate a timing error of up to 1 millisecond. The real-time solution provides IMU data paired with updated GNSS based time-stamps every 5 milliseconds. The timing jitter is reduced to a range of ±1 millisecond. Analysis of final angular rotation error and final position error from gyroscope and accelerometer readings respectively, indicate that the real-time algorithm produces a reduction in errors when the system is static, but there is no statistical evidence showing the reduction of errors from the results of the rotational tests. / GNSS / IMU integrerade navigationssystem kombinerar de positiva egenskaperna hos GNSS och IMU för optimal prestanda i noggranna navigationssystem. Detta görs med hjälp av sensorfusion, till exempel EKF. Tidssynkronisering av IMU-data med exakt GNSS-baserad tid är nödvändigt för att noggrant synkronisera de två systemen. Detta måste göras i realtid för tidskänsliga navigationsapplikationer såsom autonoma fordon. Forskningen görs i två delar. Den första delen är simulering av icke-linjär rörelse i en axel med felaktig tidsstämpling hos ett gyroskop och en accelerometer. Detta görs för att erhålla det högsta tidsfel som är acceptabelt hos ett GNSS / INS-system med hög noggrannhet. Den andra delen är skapandet av en realtidsalgoritm med ett inbyggt STM32-system med FreeRTOS som realtidskärna för en GNSSmottagare och antenn, tillsammans med en IMU-sensor. En jämförande analys av det tidssynkroniserade systemet mot ett osynkroniserat system görs baserat på de positionsfel längs en axel som produceras av gyroskopoch accelerometermätningar. Detta görs genom att utföra statiska och roterande tester med hjälp av en roterande stol.Simuleringen visar att ett noggrant GNSS / INS-system tolererar ett tidsfel på upp till 1 millisekund. Realtidslösningen ger IMU-data med tidsstämplar synkroniserade med GNSS-tid var femte millisekund. Tidsjittret reduceras till ett intervall mellan ± 1 millisekund. Analysen av det slutliga vinkelrotationsfelet och positionsfelet från gyroskopoch accelerometermätningar indikerar att realtidsalgoritmen ger ett lägre fel när systemet är statiskt. Det finns dock inga statistiska bevis för förbättringen från resultaten av rotationstesterna.

The breakthrough in mobile technology and the development of smartphones, supplied with sensing devices such as Inertial Measurement Units (IMUs), has made it possible to obtain accurate and reliable data on the angular velocity for different objects. The available technical sensors for wrist movements, such as electrogoniometers, are costly, time-consuming, and need a particular computer program to be analyzed. Therefore, there is a need to develop user-friendly risk assessment methods for wrist angular velocity measurements. This master thesis aimed to validate the accuracy of a newly developed iPhone application (App), "ErgoHandMeter," for wrist velocity in actual work tasks, by comparing the “ErgoHandMeter” to standard electrogoniometers. The project study was performed with four participants, two females and two males, from three jobs performing actual work tasks. The total angular velocity obtained by the mobile application was compared with the angular velocity data from the standard electrogoniometer. The total angular velocities obtained from the smartphone and the goniometer were computed at the 10th, 50th and 90th percentile for the four subjects. The 50th percentile of goniometer-flexion velocity (G-flex) was 7.4 ± 5.4°/s, for the goniometer-total (G-tot) 8.7 ± 6.5)°/s and for App 7.2 ± 4.9°/s. The correlation coefficient for the 50th percentile of goniometer-flexion (G-flex) parameter and smartphone application was 0.994. For the goniometer-total (G-tot) and the application, it was 0.993. In a Bland-Altman plot the mean difference between G-flex and App for the 50th percentile was -0.18 °/s and for G-tot and App was -1.54 °/s, i.e. the App was lower in average. The limit of the agreement between G-Flex and App, and G-tot and App stayed within two standard deviations. For G-Flex and App (mean+1.96SD) was 1.34 °/s, (mean-1.96SD) was -1.71 °/s, while for G-tot and App (mean+1.96SD) was 1.89 °/s, (mean-1.96SD) was -4.96 °/s, indicating an adequate agreement between the two methods. A limitation was that the included occupations were all relatively low velocity. However, in conclusion, the results indicate that the two methods agree adequately and can be used interchangeably. / Genombrottet inom mobiltekniken och utvecklingen av smarttelefoner med sensorer som t.ex. tröghetsmätningsenheter (IMU) har gjort det möjligt att få exakta och tillförlitliga uppgifter om vinkelhastigheten för olika objekt. De tillgängliga tekniska sensorerna för handledsrörelser, t.ex. elektrogoniometrar, är dyra, tidskrävande och de samplade signalerna kräver ett särskilt datorprogram för att analyseras. Det finns därför ett behov av att utveckla användarvänliga riskbedömningsmetoder för mätningar av handledens vinkelhastighet. Syftet med detta examensarbete var att validera noggrannheten hos en nyutvecklad iPhone-applikation (App), "ErgoHandMeter", för handledshastighet i verkliga arbetsuppgifter, genom att jämföra "ErgoHandMeter" med vanliga elektrogoniometrar. Projektstudien genomfördes med fyra deltagare, två kvinnor och två män, från tre yrken som utförde verkliga arbetsuppgifter. Den totala vinkelhastigheten som erhölls av mobilapplikationen jämfördes med vinkelhastighetsdata från standardelektrogoniometern. De totala vinkelhastigheterna som erhållits från smarttelefonen och goniometern beräknades vid den 10:e, 50:e och 90:e percentilen för de fyra försökspersonerna. Den 50:e percentilen för goniometer-flexionshastigheten (G-flex) var i genomsnitt 7,4°/s och för goniometertotalen (G-tot) 8,7°/s. Korrelationskoefficienten (r) för den 50:e percentilen för goniometer-flexionsparametern (G-flex) och smartphone-applikationen var 0,994. För goniometer-total (G-tot) och applikationen var r 0,993. I en Bland-Altman-plot var den genomsnittliga skillnaden mellan G-flex och appen för den 50:e percentilen -0,18°/s och för G-tot och appen -1,54°/s (App var lägre än Gon). Medelvärdet för differensen mellan G-Flex och App och G-tot och App ligger inom två standardavvikelser. För G-Flex och App (medelvärde+1,96SD) var 1,34 °/s, (medelvärde-1,96SD) var -1,71 °/s, medan för G-tot och App (medelvärde+1,96SD) var 1,89 °/s, (medelvärde-1,96SD) var -4,96 °/s. Vilket tyder på en tillräcklig överensstämmelse mellan de två metoderna. En begränsning var att de inkluderade yrkena alla hade relativt låg hastighet. Sammanfattningsvis visar dock resultaten att de två metoderna stämmer väl överens och kan användas på ett utbytbart sätt.

Mechanical overburdening is a major risk factor that provokes non-infectious claw diseases. Moreover, lameness-causing lesions often remain undetected and untreated. Therefore, prevention of claw tissue overburdening is of interest, especially by analyzing harmful effects within dairy cows’ housing environment. However, objective “on-cow” methods for bovine gait analysis are underdevel- oped. The purpose of the study was to apply an innovative mobile pressure sensor system attached at the claws to perform pedobarometric gait analysis. A further goal was the supplementation with accelerative data, generated simultaneously by use of two inertial measurement units (IMUs), attached at metatarsal level. IMU data were analyzed with an automatic step detection algorithm. Gait analysis was performed in ten dairy cows, walking and trotting on concrete flooring and rubber mats. In addition to the basic applicability of the sensor systems and with the aid of the automatic step detection algorithm for gait analysis in cows, we were able to determine the impact of the gait and flooring type on kinematic and kinetic parameters. For pressure sensor output, concrete was associated with significantly (p < 0.001) higher maximum and average pressure values and a significantly smaller contact area, compared to rubber mats. In contrast to walking, trotting led to a significantly higher force, especially under the medial claw. Further, IMU-derived parameters were significantly influenced by the gait. The described sensor systems are useful tools for detailed gait analysis in dairy cows. They allow the investigation of actors which may affect claw health negatively.

Micromobility ride-share scooters (e-scooters) have become a popular mode of transport in several major cities around the world, yet several safety and accessibility issues stem from how these scooters are operated, including sidewalk riding, unsafe parking and wrong-way riding. This thesis tackles these issues through a novel, privacy-preserving, end-to-end sensor system that employs lightweight machine learning models to provide real-time feedback to users to present unsafe scooter operation. Though this problem has been widely studied in technology startups and most of the existing solutions entail using cameras, there is an interest in academia to propose solutions that do not use cameras, as they cause privacy concerns when used in urban environments and are known to fail at night or in low-visibility conditions. Considering this, we propose a portable, cheap, and robust system that preserves privacy through the use of an Inertial Measurement Unit and radar sensors, which give only low-resolution heatmaps as outputs. Furthermore, unlike cameras, radars function even in low-visibility conditions. Using data from Inertial Measurement Unit sensors, a 1D Convolutional Neural Network is used to classify whether a scooter is being ridden on a street or on a sidewalk. Radar heatmaps are used to train a lightweight Convolutional Neural Network to classify common objects in urban environments (Lampposts, Walls, Trees, Humans, Fire Hydrants and Middle of Sidewalk). For the purpose of this research; Lamppost, Wall, Tree are considered safe parking scenarios whereas Fire Hydrant, Human and Middle of Sidewalk is considered unsafe. No public datasets exist on this topic, so two novel datasets are created. The models trained are lightweight in terms of memory and have fast inference time and are therefore suitable for edge computing, and for being deployed on a sensor system mounted on the scooter. The results achieved demonstrate the potential of this approach, though further work is needed to ensure performance for deployment in the real world. / El-sparkcyklar för korttidshyra har blivit ett populärt transportmedel i flera större städer runt om i världen, men hur dessa sparkcyklar framförs väcker flera säkerhets- och framkomlighetsfrågor, inklusive trottoarkörning, osäker parkering och körning mot enkelriktat trafik. I denna avhandling föreslår vi lösningar på dessa frågor genom ett nytt, integritetsbevarande, helomfattande sensorsystem som använder små maskininlärningsmodeller för att ge feedback i realtid till förare för att förebygga osäker körning av elsparkcyklar. Dessa problem har studerats i stor utsträckning i tekniska startups, men de flesta av de befintliga lösningarna innebär att man använder kameror. Inom akademin finns det ett intresse för att föreslå lösningar som inte använder kameror, eftersom de kan vara integritetskränkande när de används i stadsmiljöer och är kända för att ej fungera nattetid eller under förhållanden med dålig sikt. Med tanke på detta föreslår vi ett bärbart, billigt och robust system som bevarar integritet genom användning av en tröghetsmåttenhet (Inertial Measurement Unit) och radarsensorer, som endast ger lågupplösta färgdiagram som produkt. Dessutom, till skillnad från kameror, fungerar radar även under förhållanden med dålig sikt. Med hjälp av data från en tröghetsmåttenhet används ett 1D Convolutional Neural Network för att klassificera om en sparkcykel framförs på en gata eller på en trottoar. Färgdiagram från radar används för att träna ett Convolutional Neural Network för att klassificera vanligt förekommande föremål i stadsmiljöer (lyktstolpar, väggar, träd, människor, brandposter (fire hydrant) och ”mitt på trottoaren”). I detta arbete anses lyktstolpe, vägg, träd vara säkra parkeringsscenarier medan brandpost, människa och mitt på trottoaren anses vara osäkra. Det finns inga offentliga dataset för dessa ändamål, så två nya dataset skapas. Modellerna som tränas är lätta när det gäller minne och har snabb slutledningstid och är därför lämpliga för edge computing och för användning i ett sensorsystem monterat på skotern. De uppnådda resultaten visar potentialen i denna metod, även om ytterligare arbete behövs för att säkerställa prestanda för användning i den verkliga världen.

Location-based services (LBSs) have become pervasive, and the demand for these systems and services is rising. Indoor Positioning Systems (IPSs) are key to extend location-based services indoors where the Global Positioning System (GPS) is not reliable due to low signal strength and complicated signal propagation environment. Most existing IPSs either require the installation of special hardware devices or build a fingerprint map, which is expensive, time-consuming, and labor-intensive. Developments in microelectromechanical systems (MEMS) have resulted in significant advancements in the low-cost compact MARG inertial sensors, making it possible to achieve low-cost and high-accuracy IPSs. This research considers the indoor positioning problem and aims to design and develop an infrastructure-free self-contained indoor positioning and tracking system based on Pedestrian Dead Reckoning (PDR) using MEMS MARG inertial sensors. PDR-based systems rely on MARG inertial sensor measurements to estimate the current position of the object by using a previously determined position without external references. Many issues still exist in developing such systems, such as cumulative errors, high-frequency sensor noises, the gyro drift issue, magnetic distortions, etc. As the MARG sensors are inherently error-prone, the most significant challenge is how to design sensor fusion models and algorithms to accurately extract useful location-based information from individual motion and magnetic sensors. The objective of this thesis is to solve these issues and mitigate the challenges. The proposed positioning system is designed with four main modules at the system level and a dual-mode feature. Specifically, the four main modules are mode detection, step detection and moving distance estimation, heading and orientation estimation, and position estimation. To address the cumulative error issue of using low-cost inertial sensors, signal processing and sensor fusion techniques are utilized for algorithm design. Experimental evaluations show that the proposed position estimation algorithm is able to achieve high positioning accuracy at low costs for the indoor environment. / Thesis / Master of Applied Science (MASc) / With the maturity of microelectromechanical systems (MEMS) technology in recent years, Magnetic, Angular Rate, and Gravity (MARG) sensors are embedded in most smart devices. This research considers the indoor positioning problem and aims to design and develop an infrastructure-free self-contained MEMS MARG inertial sensor-based indoor positioning and tracking system with high precision. The proposed positioning system uses the Pedestrian Dead Reckoning (PDR) approach and includes four main modules at the system level with a dual-mode feature. Specifically, the four main modules are mode detection, step detection and moving distance estimation, heading and orientation estimation, and position estimation. The two modes are static mode and dynamic mode. To address the cumulative error issue of using low-cost inertial sensors, signal processing and sensor fusion techniques are utilized for algorithm design. The detection and estimation algorithms of each module are presented in the system design chapter. Experimental evaluations including trajectory results under five scenarios show that the proposed position estimation algorithm achieves a higher position accuracy than that of conventional estimation methods.