Human activity recognition based on wearable sensors’ data is quite an attractive subject due to its wide application in the fields of healthcare, wellbeing and smart environments. This research is also focussed on predictive performance comparison of machine learning algorithms for activity recognition from wearable sensors’ (MHEALTH) data while employing a comprehensive process. The framework is adapted from well-laid data science practices which addressed the data analyses requirements quite successfully. Moreover, an Analysis Tool is also developed to support this work and to make it repeatable for further work. A detailed comparative analysis is presented for five multi-class classifier algorithms on MHEALTH dataset namely, Linear Discriminant Analysis (LDA), Classification and Regression Trees (CART), Support Vector Machines (SVM), K-Nearest Neighbours (KNN) and Random Forests (RF). Beside using original MHEALTH data as input, reduced dimensionality subsets and reduced features subsets were also analysed. The comparison is made on overall accuracies, class-wise sensitivity and specificity of each algorithm, class-wise detection rate and detection prevalence in comparison to prevalence of each class, positive and negative predictive values etc. The resultant statistics have also been compared through visualizations for ease of understanding and inference. All five ML algorithms were applied for classification using the three sets of input data. Out of all five, three performed exceptionally well (SVM, KNN, RF) where RF was best with an overall accuracy of 99.9%. Although CART did not perform well as a classification algorithm, however, using it for ranking inputs was a better way of feature selection. The significant sensors using CART ranking were found to be accelerometers and gyroscopes; also confirmed through application of predictive ML algorithms. In dimensionality reduction, the subset data based on CART-selected features yielded better classification than the subset obtained from PCA technique.

No description available.

abstract: Wearable robotics has gained huge popularity in recent years due to its wide applications in rehabilitation, military, and industrial fields. The weakness of the skeletal muscles in the aging population and neurological injuries such as stroke and spinal cord injuries seriously limit the abilities of these individuals to perform daily activities. Therefore, there is an increasing attention in the development of wearable robots to assist the elderly and patients with disabilities for motion assistance and rehabilitation. In military and industrial sectors, wearable robots can increase the productivity of workers and soldiers. It is important for the wearable robots to maintain smooth interaction with the user while evolving in complex environments with minimum effort from the user. Therefore, the recognition of the user's activities such as walking or jogging in real time becomes essential to provide appropriate assistance based on the activity. This dissertation proposes two real-time human activity recognition algorithms intelligent fuzzy inference (IFI) algorithm and Amplitude omega ($A \omega$) algorithm to identify the human activities, i.e., stationary and locomotion activities. The IFI algorithm uses knee angle and ground contact forces (GCFs) measurements from four inertial measurement units (IMUs) and a pair of smart shoes. Whereas, the $A \omega$ algorithm is based on thigh angle measurements from a single IMU. This dissertation also attempts to address the problem of online tuning of virtual impedance for an assistive robot based on real-time gait and activity measurement data to personalize the assistance for different users. An automatic impedance tuning (AIT) approach is presented for a knee assistive device (KAD) in which the IFI algorithm is used for real-time activity measurements. This dissertation also proposes an adaptive oscillator method known as amplitude omega adaptive oscillator ($A\omega AO$) method for HeSA (hip exoskeleton for superior augmentation) to provide bilateral hip assistance during human locomotion activities. The $A \omega$ algorithm is integrated into the adaptive oscillator method to make the approach robust for different locomotion activities. Experiments are performed on healthy subjects to validate the efficacy of the human activities recognition algorithms and control strategies proposed in this dissertation. Both the activity recognition algorithms exhibited higher classification accuracy with less update time. The results of AIT demonstrated that the KAD assistive torque was smoother and EMG signal of Vastus Medialis is reduced, compared to constant impedance and finite state machine approaches. The $A\omega AO$ method showed real-time learning of the locomotion activities signals for three healthy subjects while wearing HeSA. To understand the influence of the assistive devices on the inherent dynamic gait stability of the human, stability analysis is performed. For this, the stability metrics derived from dynamical systems theory are used to evaluate unilateral knee assistance applied to the healthy participants. / Dissertation/Thesis / Doctoral Dissertation Aerospace Engineering 2018

Le travail présenté dans ce mémoire porte sur le développement d'un système de surveillance ambulatoire pour la détection de la dyskinésie induite par la Levodopa (LID) chez les patients de la maladie de Parkinson (PD). Le système est composé d’unités de mesure inertielle (IMUs) qui recueillent des signaux de mouvement chez des sujets sains et des patients parkinsoniens. Des méthodes différentes sont évaluées pour la détection de LID avec et sans classification des activités. Les données recueillies auprès des sujets sains sont utilisées pour concevoir un classificateur d'activité fiable. Par la suite, un algorithme qui effectue la classification des activités et la détection de la dyskinésie sur les données recueillies auprès de des patients parkinsoniens est proposé. Une nouvelle approche basée sur l'analyse de réseau complexe est également explorée et présente des résultats intéressants. Les méthodes de traitement développées ont été intégrées dans une plateforme complète d’analyse nommée PARADYSE. / The work presented in this thesis is concerned with the development of an ambulatory monitoring system for the detection of Levodopa Induced Dyskinesia (LID) in Parkinson’s disease (PD) patients. The system is composed of Inertial Measurement Units (IMUs) that collect movement signals from healthy individuals and PD patients. Different methods are evaluated which consist of LID detection with and without activity classification. Data collected from healthy individuals is used to design a reliable activity classifier. Following that, an algorithm that performs activity classification and dyskinesia detection on data collected from PD patients is tested. A new approach based on complex network analysis is also explored and presents interesting results. The evaluated analysis methods are incorporated into a platform PARADYSE in order to further advance the system’s capabilities.

This work explores the blending of e-textile technology with the porous electrode of polymer electrolyte membrane fuel cells (PEMFCs) and with smart wound patches to allow monitoring and in-situ diagnostics. This work includes contributions to understanding water transport and conductivity in the carbon cloth gas diffusion layer (GDL), and further developing thread-based relative humidity (RH) and temperature sensors, which can be sewn on a cloth GDL in PEMFCs. We also explore the application of the developed RH and temperature sensors in wearable biomonitoring. First, an experimental prototype is developed for evaluating water transport, thermal conductivity and electrical conductivity of carbon cloth GDLs under different hydrophobic coatings and compressions. Second, we demonstrate the addition of external threads to the carbon cloth GDL to (1) facilitate water transport and (2) measure local RH and temperature with a minimal impact on the physical, microstructural and transport properties of the GDL. We illustrate the roll-to-roll process for fabricating RH and temperature sensors by dip-coating commodity threads into a carbon nanotubes (CNTs) suspension. The thread-based sensors response to RH and temperature in the working environment of PEMFCs is investigated. As a proof-of-concept, the local temperature of carbon cloth GDL is monitored in an ex-situ experiment. Finally, we optimized the coating parameters (e.g. CNTs concentration, surfactant concentration and a number of dipping) for the thread-based sensors. The response of the thread-based sensors in room conditions is evaluated and shows a linear resistance decrease to temperature and a quadratic resistance increase to RH. We also evaluated the biocompatibility of the sensors by performing cell cytotoxicity and studying wound healing in an animal model. The novel thread-based sensors are not only applicable for textile electrochemical devices but also, show a promising future in wearable biomonitoring applications. / Graduate

The contact patterns among individuals can significantly affect the progress of an infectious outbreak within a population. Gathering data about these interaction and mixing patterns is essential to assess computational modeling of infectious diseases. Various self-report approaches have been designed in different studies to collect data about contact rates and patterns. Recent advances in sensing technology provide researchers with a bilateral automated data collection devices to facilitate contact gathering overcoming the disadvantages of previous approaches. In this study, a novel unilateral wearable sensing architecture has been proposed that overcome the limitations of the bi-lateral sensing. Our unilateral wearable sensing system gather contact data using hybrid sensor arrays embedded in wearable shirt. A smartphone application has been used to transfer the collected sensors data to the cloud and apply deep learning model to estimate the number of human contacts and the results are stored in the cloud database. The deep learning model has been developed on the hand labelled data over multiple experiments. This model has been tested and evaluated, and these results were reported in the study. Sensitivity analysis has been performed to choose the most suitable image resolution and format for the model to estimate contacts and to analyze the model's consumption of computer resources.

Stress and overload are strongly associated with unsafe behaviour, which motivated various studies to detect them automatically in workplaces. This study aims to advance safety research by developing a data-driven stress and overload detection method. An unsupervised deep learning-based anomaly detection method is developed to detect stress. The proposed method performs with convolutional neural network encoder-decoder and long short-term memory equipped with an attention layer. Data from a field experiment with 18 participants was used to train and test the developed method. The field experiment was designed to include a pre-defined sequence of activities triggering mental and physical stress, while a wristband biosensor was used to collect physiological signals. The collected contextual and physiological data were pre-processed and then resampled into correlation matrices of 14 features. Correlation matrices are used as an input to the unsupervised Deep Learning (DL) based anomaly detection method. The developed method is validated, offering accuracy and F-measures close to 0.98. The technique employed captures the input data attributes correlation, promoting higher interpretability of the DL method for easier comprehension. Over-reliance on uncertain absolute truth, the need for a high number of training samples, and the requirement of a threshold for detecting anomalies are identified as shortcomings of the proposed method. To overcome these shortcomings, an Adaptive Neuro-Fuzzy Inference System (ANFIS) was designed and developed. While the ANFIS method did not improve the overall accuracy, it outperformed the DL-based method in detecting anomalies precisely. The overall performance of the ANFIS method is better than the DL-based method for the anomalous class, and the method results in lower false alarms. However, the DL-based method is suitable for circumstances where false alarms are tolerated. / Dissertation / Doctor of Philosophy (PhD)

Ergonomic assessments to determine risks of work-related musculoskeletal disorders as well as to compare designs of work-tasks and workstations, are imperative for high sustainability and productivity in any given industry. Hence, assessment tools that can objectively and reliably capture postures, joint angles and muscle activities play very important role in properly determining risks relating to work and various tasks. The introduction of direct measurement instruments/tools has been helping and continue to help improve upon observational assessment methods to attain objectivity and reliability. This project aimed at contributing to future improvements of industrial risk assessment measures in ergonomics by identifying and testing direct measurement instruments/tools that can enhance observational risk assessment methods and introduce a new way of signal processing, hence, reducing assessment time while increasing objectivity and reliability. Several candidate instruments were identified and out of the identified, ten were selected as potential candidates. Two out of the ten, Wergonic and ErgoHandMeter were then selected and tested on common observation risk assessment factors that could be measured and answers provided directly or by analyses. The Wergonic instrument was modified to enhance its measuring capability from one fully and partially two factors to six factors. New algorithms were also employed to analyse measurements of the ErgoHandMeter in order to answer questions regarding repetitive movements. The two instruments tested, are able to measure and provide results for six commonly and one rarely assessed biomechanical risk factors. By combining selected potential candidates, many of the commonly targeted biomechanical risk factors in observational instruments can be measured by the selected direct measurement instruments. However, some factors especially force measurement remain a challenge for measuring by direct wearable sensor instruments.

A growing number of factors, including costs, technological advancements, ageing populations, and medical errors, are leading industrialized countries to invest in research on alternative solutions to improve their health-care systems and increase patients’ quality of life. Personal health systems (PHS) examplify the use of information and communication technologies that enable a paradigm shift from the traditional hospital-centered healthcare delivery model toward a preventive and person-centered approach. PHS offer the means to monitor a patient’s health using wearable, portable or implantable systems that offer ubiquitous, unobtrusive biodata acquisition, allowing remote monitoring of treatment and access to the patient’s status. Electrical bioimpedance (EBI) technology is non-invasive, quick and relatively affordable technique that can be used for assessing and monitoring different health conditions, e.g., body composition assessments for nutrition. When combined with state-of-the-art advances in sensors and textiles, EBI technologies are fostering the implementation of wearable bioimpedance monitors that use functional garments for personalized healthcare applications. This research work is focused on the development of wearable EBI-based monitoring systems for ubiquitous health monitoring applications. The monitoring systems are built upon portable monitoring instrumentation and custom-made textile electrode garments. Portable EBI-based monitors have been developed using the latest material technology and advances in system-on-chip technology. For instance, a portable EBI spectrometer has been validated against a commercial spectrometer for total body composition assessment using functional textile electrode garments. The development of wearable EBI-based monitoring units using functional garments and dry textile electrodes for body composition assessment and respiratory monitoring has been shown to be a feasible approach. The availability of these measurement systems indicates progress toward the real implementation of personalized healthcare systems. / <p>QC 20170517</p>

This thesis in the field of microdata analysis aims to introduce dose optimizing algorithms for the pharmacological management of Parkinson’s disease (PD). PD is a neurodegenerative disease that mostly affects the motor functions of the patients and it is characterized as a movement disorder. The core symptoms of PD are: bradykinesia, postural instability, rigidity, and tremor. There is no cure for PD and the use of levodopa to manage the core symptoms is considered the gold standard. However, long term use of levodopa causes reduced medication efficacy, and side effects, such as dyskinesia, which can also be attributed to overmedication. When that happens precise individualized dosing schedules are required. The goal of this thesis is to examine if algorithmic methods can be used to find dosing schedules that treat PD symptoms and minimize manifestation of side effects. Data from three different sources were used for that purpose: data from a clinical study in Uppsala University hospital in 2015, patient admission chart data from Uppsala University hospital during 2011-2015, and data from a clinical study in Gothenburg University during 2016-2017. The data were used to develop the methods and evaluate the performance of the proposed algorithms.The first algorithm that was developed was a sensor-based method that derives objective measurements (ratings) of PD motor states. The construction of the sensor index was based on subjective ratings of patients’ motor functions made by three movement disorder experts. This sensor-based method was used when deriving algorithmic dosing schedules. Afterwards, a method that uses medication information and ratings of the patients’ motor states to fit individual patient models was developed. This method uses mathematical optimization to individualize specific parameters of dose-effects models for levodopa intake, through minimizing the distance between motor state ratings and dose-effect curves. Finally, two different dose optimization algorithms were developed and evaluated, that had as input the individual patient models. The first algorithm was specific to continuous infusion of levodopa treatment, where the patient’s state was set to a specific target value and the algorithm made dosing adjustments to keep that patients motor functions on that state. The second algorithm concerned oral administration of microtables of levodopa. The ambition with this algorithm was that the suggested doses would find the right balance between treating the core symptoms of PD and, at the same time, minimizing the side effects of long term levodopa use, mainly dyskinesia. Motor state ratings for this study were obtained through the sensor index. Both algorithms followed a principle of deriving a morning dose and a maintenance dose for the patients, with maintenance dose being an infusion rate for the first algorithm, and oral administration doses at specific time points for the second algorithm.The results showed that the sensor-based index had good test-retest reliability, sensitivity to levodopa treatment, and ability to make predictions in unseen parts of the dataset. The dosing algorithm for continuous infusion of levodopa had a good ability to suggest an optimal infusion rating for the patients, but consistently suggested lower morning dose than what the treating personnel prescribed. The dosing algorithm for oral administration of levodopa showed great agreement with the treating personnel’s prescriptions, both in terms of morning and maintenance dose. Moreover, when evaluating the oral medication algorithm, it was clear that the sensor index ratings could be used for building patient specific models.