"Precise and accurate localization and motion classification is an emerging fundamental areas for scientific research and engineering developments. Such science and technology began from the broad out door area applications, and gradually grew into smaller and more complicated in-door area and more recently it is proceeding into in-body area networking for medical applications. Localization and motion classification technologies have their own specific challenges depending on the application and environment, which are left for scientists and engineers to overcome. One major challenge is that location estimation and motion classification often use hand-held devices or wearable sensors. Such devices and sensors usually work in indoor, near body environments and the human object has certain effects on the measurements. In that situation, existing mathematical models for general environments are no longer accurate and new models and analytical approaches are required to deal with the human body effects. This has opened opportunities for researchers to tackle a number of demanding problems. This dissertation focuses on three novel problems in localization and motion classification using radio propagation (RF) modeling, in and around the human body. (1) We develop an empirical Time-of-Arrival (TOA) ranging error model for radio propagation from body-mounted sensors to external access points, for human body tracking in indoor environment. This model reflects the effects of human angular motion on TOA ranging estimation, which enables accurate analysis for conventional TOA-based human tracking systems. (2) We use empirical data collected from a RF connection between a pair of body-mounted sensors to classify seven frequently appeared human body motions. This RF based classification approach has enabled health monitoring applications for first responders, hospital patient, and elderly care centers and in most of the situations it can replace the costly video base monitoring systems. (3) We use radio propagation models from body-mounted sensor to medical implants and the moving pattern of micro-robots inside the body to analyze the accuracy of hybrid localization inside the human body. This analysis demonstrates the feasibility of millimeter level of accurate localization inside the human body, which opens up possibilities for 3D reconstruction of the interior of human GI tract."

No description available.

Currently, most long-term human motion classification and prediction tasks are driven by spatio-temporal data of the human trunk. In addition, data with multiple modalities can change idiosyncratically with human motion, such as electromyography (EMG) of specific muscles and respiratory rhythm. On the other hand, progress in Artificial Intelligence research on the collaborative understanding of image, video, audio, and semantics mainly relies on MultiModal Machine Learning (MMML). This work explores human motion classification strategies with multi-modality information using MMML. The research is conducted using the Unige-Maastricht Dance dataset. Attention-based Deep Learning architectures are proposed for modal fusion on three levels: 1) feature fusion by Component Attention Network (CANet); 2) model fusion by fusing Graph Convolution Network (GCN) with CANet innovatively; 3) and late fusion by a simple voting. These all successfully exceed the benchmark of single motion modality. Moreover, the effect of each modality in each fusion method is analyzed by comprehensive comparison experiments. Finally, statistical analysis and visualization of the attention scores are performed to assist the distillation of the most informative temporal/component cues characterizing two qualities of motion. / För närvarande drivs uppgifter som långsiktig klassificering och förutsägelse av mänskliga rörelser av spatiotemporala data från människans bål. Dessutom kan data från flera olika modaliteter förändras idiosynkratiskt med mänsklig rörelse, t.ex. elektromyografi (EMG) av specifika muskler och andningsrytm. Å andra sidan bygger forskning inom artificiell intelligens för samtidig förståelse av bild, video, ljud och semantik huvudsakligen på multimodal maskininlärning (MMML). I det här arbetet undersöks strategier för klassificering av mänskliga rörelser med multimodal information med hjälp av MMML. Forskningen utförs med hjälp av Unige-Maastricht Dance dataset. Uppmärksamhetsbaserade djupinlärningsarkitekturer föreslås för modal fusion på tre nivåer: 1) funktionsfusion genom Component Attention Network (CANet), 2) modellfusion genom en innovativ fusion av Graph Convolution Network (GCN) med CANet, 3) och sen fusion genom en enkel omröstning. Alla dessa överträffar riktmärket med en enda rörelsemodalitet. Dessutom analyseras effekten av varje modalitet i varje fusionsmetod genom omfattande jämförelseexperiment. Slutligen genomförs en statistisk analys och visualiseras av uppmärksamhetsvärdena för att hjälpa till att hitta de mest informativa temporala signaler eller komponentsignaler som kännetecknar två typer av rörelse.