The purpose of this thesis was to determine the proportion of participants enrolled in a community-based weight loss program that would accept and use a wearable device (Fitbit) if included as part of the program. A sample of 526 newly enrolled, adult, female weight loss program participants (BMI ≥ 30 kg/m2 ) were recruited. Participants were randomized to either a Fitbit experimental condition or no-Fitbit control condition, and received emailed information on program features. The experimental condition email also included a free Fitbit offer. The full sample (n=526) was 44±12.6 years old with a BMI of 37±6.2 kg/m2. The proportion of experimental sample (n=266) that accepted and synced was 50% and 23%, respectively. Twenty-two participants in the control condition (8%) also independently obtained and synced a Fitbit. Ninety-nine percent passively declined (did not respond to request for Fitbit color and size information). Those that declined were older (46±13.4 vs. 42±11.3 years of age, p=.001) and weighed less (214±38.9lbs. vs. 231±41.3lbs., p=.01) than those who accepted. Those in the experimental sample who synced were younger (42±10.0 vs. 45±13.2 years of age, p=.012), and weighed more (237±45.2lbs. vs. 217±38.1lbs., p=.002) than those who accepted but did not sync. This thesis provides preliminary support that 23% of participants will accept and sync a free wearable device. These data can be used for decision making, combined with effectiveness and cost data, and research on wearable activity trackers and community, incentive, and web-based weight loss. / Master of Science

Wearable activity classifiers, so far, have been able to perform well with simple activities, strictly-scripted activities, and application-specific activities. In addition, current classification systems suffer from using impractical tight-fitting sensor networks, or only use one loose-fitting sensor node that cannot capture much movement information (e.g., smartphone sensors and wrist-worn sensors). These classifiers either do not address the bigger picture of making activity recognition more practical and being able to recognize more complex and naturalistic activities, or try to address this issue but still perform poorly on many fronts. This dissertation works toward having practical, in-the-wild, and reusable wearable activity classifiers by taking several steps that include the four following main contributions. The dissertation starts by quantifying users' needs and expectations from wearable activity classifiers to set a framework for designing ideal wearable activity classifiers. Data collected from user studies and interviews is gathered and analyzed, then several conclusions are made to set a framework of essential characteristics that ideal wearable activity classification systems should have. Afterwards, this dissertation introduces a group of datasets that can be used to benchmark different types of activity classifiers and can accommodate for a variety of goals. These datasets help comparing different algorithms in activity classification to assess their performance under various circumstances and with different types of activities. The third main contribution consists of developing a technique that can classify complex activities with wide variations. Testing this technique shows that it is able to accurately classify eight complex daily-life activities with wide variations at an accuracy rate of 93.33%, significantly outperforming the state-of-the-art. This technique is a step forward toward classifying real-life natural activities performed in an environment that allows for wide variations within the activity. Finally, this dissertation introduces a method that can be used on top of any activity classifier that allows access to its matching scores in order to improve its classification accuracy. Testing this method shows that it improves classification results by 11.86% and outperforms the state-of-the-art, therefore taking a step forward toward having reusable activity classification techniques that can be used across users, sensor domains, garments, and applications. / Ph. D. / Wearable activity classifiers are wearable systems that can recognize human activities. These systems are needed in many applications. Nowadays, they are mainly used for fitness purposes – e.g., Fitbits and Apple Watches – and in gaming consoles – e.g., Microsoft Kinect. However, these systems are still far from being ideal. They still miss many characteristics that would make them practical and usable for different purposes, such as in medical applications, industrial applications, and other types of applications where recognizing human activities can be useful. This dissertation works toward having practical wearable activity classifiers that can be reused for different purposes in real-life scenarios. Four contributions are introduced in this dissertation. The dissertation starts by quantifying users’ needs and expectations from wearable activity classifiers and sets a framework for designing them. Afterward, this dissertation introduces a group of datasets that can be used to benchmark and compare different types of activity classifiers. The third main contribution consists of a technique that enables activity classifiers to recognize complex activities having a wide range of variations within each activity. Finally, this dissertation introduces a method that can be used to improve the recognition accuracy of activity classifiers.

Upper Limb work-related musculoskeletal disorders (WMSDs) present a significant health risk to industrial workers. To address this, rigid-body exoskeletons have been widely used in industrial settings to mitigate these risks while exosuits offer advantages such as reduced weight, lower inertia, and no need for precise joint alignment, However, they remain in the early stages of development, especially for reducing muscular effort in repetitive and forceful tasks like heavy lifting and overhead work. This study introduces a multiple degrees-of-freedom cable-driven upper limb bilateral exosuit for human power augmentation. Two control schemes were developed and compared: an IMU based controller, and a myoelectric controller to compensate for joint torque exerted by the wearer. The results of preliminary experiments showed a substantial reduction in muscular effort with the exosuit's assistance, with the myoelectric control scheme exhibiting reduced operational delay. In parallel, the neuromusculoskeletal modeling and simulator (NMMS) has been widely applied in various fields. Most of the research works implements the PD-based internal model of human’s central nervous system to simulate the generated muscle activation. However, the PD-based internal models in recent works are tuned by the empirical data which requires empirical data from human subject experiments. In this dissertation, an off-policy DRL algorithm, Deep Deterministic Policy Gradient was implemented to tune the PD-based internal model of human’s central nervous system. Compared to the conventional approaches, the DRL-based auto-tuner can learn the optimal policy through trial-and-error which doesn’t require human subject experiment and empirical data. The experiment this work showed promising results of this DRL-based auto-tuner for internal-model of human’s central nervous system.

During walking, arm swing helps maintain postural balance and stability, but it does not aid in body propulsion. Human bipedal locomotion makes the upper limbs passively swing without engaging much upper limb muscle force or effort. The central idea of this thesis is to capture the kinetic energy of the arm swing during walking and transfer it to the lower limbs via a wearable exosuit to reduce the lower limb muscle efforts during walking. The Arm-Leg Interactive Coupling Exosuit (ALICE) is designed with cable-pulley system to harness shoulder and elbow movements to support the hips and ankles during walking. ALICE employs two coupling methods, shoulder-hip coupling and elbow-ankle coupling. The shoulder is coupled with the hip contralaterally, while the elbow is coupled with the ankle ipsilaterally to match the natural walking pattern. Therefore, shoulder flexion results in hip flexion, and elbow flexion initiates plantarflexion of the ankle during toe-off. The proposed concept was validated through human subject experiments involving 15 healthy young adults who walked on a treadmill for 5 minutes with and without the device. Walking kinematics, muscle activity, foot pressure, and metabolic cost were recorded to compare differences in walking biomechanics and energetics between three conditions - BL, S1 (device worn but disengaged) and S2 (device worn and engaged). A repeated measures analysis of variance (ANOVA) followed by post hoc analysis was used to identify the effects of shoulder-hip coupling, elbow-ankle coupling, and their combined effects. The results indicate that the proposed concepts indeed generate the expected outcomes of reducing lower limb muscle activity in exchange for the increased effort of upper limb muscles.

Yes / Many children are not engaging in sufficient physical activity and there are substantial between-children physical activity inequalities. In addition to their primary role as educators, teachers are often regarded as being well-placed to make vital contributions to inclusive visions of physical activity promotion. With the dramatic increase in popularity of wearable technologies for physical activity promotion in recent years, there is a need to better understand teachers' perspectives about using such devices, and the data they produce, to support physical activity promotion in schools. Method: Semi-structured interviews were conducted with 26 UK-based primary school teachers, exploring their responses to children's physical activity data and their views about using wearable technologies during the school day. Interview discussions were facilitated by an elicitation technique whereby participants were presented with graphs illustrating children's in-school physical activity obtained from secondary wearable technology data. Interview transcripts were thematically analyzed. Results: Most teachers spoke positively about the use of wearable technologies specifically designed for school use, highlighting potential benefits and considerations. Many teachers were able to understand and critically interpret data showing unequal physical activity patterns both within-and between-schools. Being presented with the data prompted teachers to provide explanations about observable patterns, emotional reactions-particularly about inequalities-and express motivations to change the current situations in schools. Conclusion: These findings suggest that primary school teachers in the UK are open to integrating wearable technology for measuring children's physical activity into their practices and can interpret the data produced by such devices. Visual representations of physical activity elicited strong responses and thus could be used when working with teachers as an effective trigger to inform school practices and policies seeking to address in-school physical inactivity and inequalities.

2025 / The skin, with its high density of specialized neurons, provides a rich platform for discrete communication through haptic feedback technologies. However, current manufacturing techniques for haptic devices are labor-intensive and require significant technical expertise, limiting accessibility and broader adoption. Existing processes involve specialized PCB software for circuit design, followed by multi-step integration into a soft polymer matrix, resulting in prolonged lead times and limited design flexibility. Furthermore, these devices typically rely on external computing units for controlling tactile patterns and intensity, often decoupling the two. This work proposes fabrication strategies ranging from benchtop 3D printing to hybrid techniques that integrate innovative materials with intuitive interfaces, enabling customizable and accessible haptic devices. We aim to create wearable haptic devices with direct, human-in-the-loop customization of haptic cues. To simplify the creation of haptic feedback devices, we developed a toolkit comprising five wireless, wearable haptic modules that deliver the three most common tactile sensations: vibrotactile, skin-stretch, and probing. These customizable modules can operate individually or together to create multimodal haptic experiences, serving as a platform for rapid prototyping tactile displays. However, despite their accessibility and ease of assembly, the modules remain bulky, rigid, and limited in customization, relying on an off-board computer and technical expertise to function. To create truly body-compliant stretchable haptic electronics, we developed a 3D printed liquid metal (LM) emulsion for wiring that sustains high strains while maintaining electrical connectivity. To fabricate stretchable electronics, the LM emulsion was integrated into a soft polymer matrix through multi-material 3D printing, with manually placed off-the-shelf electronics. The LM emulsion is not conductive upon printing but can be render highly conductive with a single axial strain at low stress (< 0.3 MPa), resulting in activation stresses that are an order of magnitude lower than previous work. The LM emulsion also exhibits a maximum conductivity that is more than 300% higher than that of similar previous work. Its high conductivity and durability under strain make it ideal for stretchable electronics. To integrate the LM emulsion into stretchable electronics, we developed a computer aided fabrication strategy that streamlined the design and production of haptic devices. First, we created an intuitive graphical user interface (GUI) for sketching haptic devices, compatible with direct ink writing. Next, we developed an algorithm to convert circuit schematics into 3D printing commands. This strategy combines direct ink writing with automated pick-and-place of electronics in a single fabrication step. Using this process, we fabricated a wireless, self-powered tactile display comprising a haptic input device and a haptic output device. Together, these devices enable immersive human-to-human interactions by mapping real-time pressure patterns through the input device and generating proportional vibrotactile feedback with the output device. This approach represents a significant step toward enabling rapid prototyping of both haptic devices and haptic experiences. / 2026-02-07T00:00:00Z

2025 / Movement is a window into health and disease. Advanced laboratory tools like force plates have significantly advanced movement science; however, their inaccessibility in clinical settings due to cost, time, and expertise prevents clinicians from directly utilizing the measurement, diagnostic, and predictive insights these tools offer. Consequently, current approaches to the neuromotor rehabilitation of highly prevalent neurological conditions – stroke and Parkinson's disease – have limited ability to provide optimal, personalized care. Despite clinicians' desire for personalized, targeted interventions that can enhance patient outcomes, the current movement measurement gap in clinical settings often leads to suboptimal care. Given the heterogeneity of movement impairments in these patient populations, there is a pressing need for clinically accessible movement measurement tools to guide tailored treatments. This dissertation presents the development and evaluation of novel movement assessment algorithms that utilize a minimal set of wearable inertial measurement unit (IMU) sensors to accurately estimate clinically relevant gait metrics in patients with stroke or Parkinson’s disease. Additionally, using a single thigh-mounted IMU, we identified unique movement phenotypes that show promise as diagnostic biomarkers of neurological disease and gait impairment, as well as predictive biomarkers of treatment response. Together, these findings advance the translation of movement science into clinical practice, unlocking the practical use of wearable sensors by clinicians. By facilitating in-clinic estimation of gait metrics and identification of movement phenotypes, this work supports advanced clinical-decision making, including the development of targeted treatment plans and the prediction of recovery outcomes, thereby advancing movement as a true window into health and disease. / 2026-01-29T00:00:00Z

Internet of Things is expanding to the areas such as healthcare, home management, industrial, agriculture, and becoming pervasive in our life, resulting in improved efficiency, accuracy and economic benefits. Smart sensors with embedded interfacing integrated circuits (ICs) are important enablers, hence, variety of smart sensors are required. However, each type of sensor requires specific interfacing chips, which divides the huge market of sensors’ interface chips into lots of niche markets, resulting in high develop cost and long time-to-market period for each type. Software defined sensor is regarded as a promising solution, which is expected to use a flexible interface platform to cover different sensors, deliver specificity through software programming, and integrate easily into the Internet of Things. In this work, research is carried out on the design and implementations of ultra low power analog interface circuits toward software defined sensors for healthcare services based on Internet of Things.    This thesis first explores architectures and circuit techniques for energy-efficient and flexible analog to digital conversion. A time-spreading digital calibration, to calibrate the errors due to finite gain and capacitor mismatch in multi-bit/stage pipelined converters, is developed with short convergence time. The effectiveness of the proposed technique is demonstrated with intensive simulations. Two novel circuit level techniques, which can be combined with digital calibration techniques to further improve the energy efficiency of the converters, are also presented. One is the Common-Mode-Sensing-and-Input-Interchanging (CSII) operational-transconductance-amplifier (OTA) sharing technique to enable eliminating potential memory effects. The other is a workload-balanced multiplying digital-to-analog converter (MDAC) architecture to improve the settling efficiency of a high linear multi-bit stage. Two prototype converters have been designed and fabricated in 0.13 μm CMOS technology. The first one is a 14 bit 50 MS/s digital calibrated pipelined analog to digital converter that employs the workload-balanced MDAC architecture and time-spreading digital calibration technique to achieve improved power-linearity tradeoff. The second one is a 1.2 V 12 bit 5~45 MS/s speed and power-scalable ADC incorporating the CSII OTA-sharing technique, sample-and-hold-amplifier-free topology and adjustable current bias of the building blocks to minimize the power consumption. The detailed measurement results of both converters are reported and deliver the experimental verification of the proposed techniques.     Secondly, this research investigates ultra-low-power analog front-end circuits providing programmability and being suitable for different types of sensors. A pulse-width- -modulation-based architecture with a folded reference is proposed and proven in a 0.18 μm technology to achieve high sensitivity and enlarged dynamic range when sensing the weak current signals. A 8-channel bio-electric sensing front-end, fabricated in a 0.35 μm CMOS technology is also presented that achieves an input impedance of 1 GΩ, input referred noise of 0.97 Vrms and common mode rejection ratio of 114 dB. With the programmable gain and cut-off frequency, the front-end can be configured to monitor for long-term a variety of bio-electric signals, such as electrooculogram (EOG), electromyogram (EMG), electroencephalogram (EEG) and electrocardiogram (ECG) signals. The proposed front-end is integrated with dry electrodes, a microprocessor and wireless link to build a battery powered E-patch for long-term and continuous monitoring. In-vivo test results with dry electrodes in the field trials of sitting, standing, walking and running slowly, show that the quality of ECG signal sensed by the E-patch satisfies the requirements for preventive cardiac care.    Finally, a wireless multimodal bio-electric sensor system is presented. Enabled by a customized flexible mixed-signal system on chip (SoC), this bio-electric sensor system is able to be configured for ECG/EMG/EEG recording, bio-impedance sensing, weak current stimulation, and other promising functions with biofeedback. The customized SoC, fabricated in a 0.18 μm CMOS technology, integrates a tunable analog front-end, a 10 bit ADC, a 14 bit sigma-delta digital to current converter, a 12 bit digital to voltage converter, a digital accelerator for wavelet transformation and data compression, and a serial communication protocol. Measurement results indicate that the SoC could support the versatile bio-electric sensor to operate in various applications according to specific requirements. / <p>QC 20151221</p>

Recent advances in camera technology have made it possible to build a comfortable, wearable system which can capture the scene in front of the user throughout the day. Products based on this technology, such as GoPro and Google Glass, have generated substantial interest. In this thesis, I present my work on egocentric vision, which leverages wearable camera technology and provides a new line of attack on classical computer vision problems such as object categorization and activity recognition. The dominant paradigm for object and activity recognition over the last decade has been based on using the web. In this paradigm, in order to learn a model for an object category like coffee jar, various images of that object type are fetched from the web (e.g. through Google image search), features are extracted and then classifiers are learned. This paradigm has led to great advances in the field and has produced state-of-the-art results for object recognition. However, it has two main shortcomings: a) objects on the web appear in isolation and they miss the context of daily usage; and b) web data does not represent what we see every day. In this thesis, I demonstrate that egocentric vision can address these limitations as an alternative paradigm. I will demonstrate that contextual cues and the actions of a user can be exploited in an egocentric vision system to learn models of objects under very weak supervision. In addition, I will show that measurements of a subject's gaze during object manipulation tasks can provide novel feature representations to support activity recognition. Moving beyond surface-level categorization, I will showcase a method for automatically discovering object state changes during actions, and an approach to building descriptive models of social interactions between groups of individuals. These new capabilities for egocentric video analysis will enable new applications in life logging, elder care, human-robot interaction, developmental screening, augmented reality and social media.

How can vibrotactile stimuli be used to create a technology-mediated somatic learning experience? This question motivates this practice-based research, which explores how the Feldenkrais Method and cognate neuroscience research can be applied to technology design. Supported by somaesthetic philosophy, soma-based design theories, and a critical acknowledgement of the socially-inflected body, the research develops a systematic method grounded in first- and third-person accounts of embodied experience to inform the creation and evaluation of design of Haplós, a wearable, user-customisable, remote-controlled technology that plays methodically composed vibrotactile patterns on the skin in order to facilitate body awareness—the major outcome of this research and a significant contribution to soma-based creative work. The research also contributes to design theory and somatic practice by developing the notion of a somatic learning affordance, which emerged during course of the research and which describes the capacity of a material object to facilitate somatic learning. Two interdisciplinary collaborations involving Haplós contribute to additional fields and disciplines. In partnership with experimental psychologists, Haplós was used in a randomised controlled study that contributes to cognitive psychology by showing that vibrotactile compositions can reduce, with statistical significance, intrusive food-related thoughts. Haplós was also used in Bisensorial, an award-winning, collaboratively developed proof-of-concept of a neuroadaptive vibroacoustic therapeutic device that uses music and vibrotactile stimuli to induce desired mental states. Finally, this research contributes to cognitive science and embodied philosophy by advancing a neuroscientific understanding of vibrotactile somaesthetics, a novel extension of somaesthetic philosophy.