Wi-Fi access points have been widely deployed in homes, offices, and public spaces. Some APs allow users to adjust the antenna orientation to improve communication performance by optimizing antenna polarization. However, it is difficult for non-expert users to determine the optimal orientation, and users often leave the antenna orientation in ineffective positions. To address this issue, we developed a mechanical Wi-Fi antenna device capable of automatically tuning its orientation. Experimental results show that antenna orientation could cause a throughput variation of approximately 70 Mbps under line-of-sight conditions. Furthermore, Bayesian optimization identified better configurations than random search, demonstrating its effectiveness for orientation tuning.

Wireless sensing is an emerging technology with a wide range of applications, but most existing systems capture only the motion of a single point, such as in respiration monitoring. This limitation is critical for tasks requiring multi-point data, such as respiratory volume measurement, where different body points provide distinct information, and a single point cannot represent them all. In this paper, we propose LiSen, a smartphone-integrated LiDAR system for multi-point wireless sensing, and demonstrate its contact-free capability for measuring respiratory volume. LiSen uses smartphone LiDAR to track multiple chest and abdominal points, enabling the first ranging-based spirometer system that captures the full volume curve without new-user calibration. We leverage the unique feature of multi-point sensing to address challenges such as body interference, diverse breathing patterns, and pressure differences. Tests with 35 examinees show that LiSen accurately estimates both instantaneous forced expiratory and inspiratory volume, achieving mean absolute errors below 0.24 L and 0.30 L, respectively, and an 8.93% error for four common pulmonary function indices.

Decentralized federated learning (FL) is a promising approach for training machine learning models on sensor networks, Internet of Things (IoT) devices, and other edge systems where no central server exists. While federated learning offers advantages such as preserving data privacy, it often suffers from non-independent and identically distributed (IID) data distributions across devices, which cause significant performance degradation. This issue is particularly severe when directly optimizing model parameters, because neural network training is inherently non-convex and standard convergence guarantees for convex optimization do not apply. Unlike existing decentralized FL methods that primarily operate in parameter space, we propose federated function-space alternating direction method of multipliers (FedF-ADMM). FedF-ADMM exploits the convexity of loss functionals within function space to derive alternating direction method of multipliers (ADMM)-based update directions, which are subsequently projected onto the parameter space via knowledge distillation. We further introduce a stabilization coefficient to enhance robustness under severe non-IID settings and analyze its behavior from a control-theoretic perspective by interpreting it as a proportional-integral (PI) term. Experiments under challenging non-IID scenarios, including settings where each device has data from only a single label, demonstrate that FedF-ADMM achieves faster and more stable convergence than existing decentralized FL methods, while attaining higher accuracy and better consensus among devices.

Micro-mobility has become a vital means of transportation in recent years, however, it has also resulted in a rise in traffic incidents. Timely tracking and predicting riders' maneuvers hold the potential to ensure active protection and allow for sufficient time to avert accidents by issuing timely warnings and interventions. We contend that the rider's head dynamics can provide valuable information regarding their subsequent maneuvers. Riders' traveling habits, however diverse, not to mention the rapidly varying riding environment. The above factors contribute to significant disruptions in the data source, and various micro-mobility forms further exacerbate the issue. We accordingly present HeadMon+, which predicts the rider's subsequent maneuver by examining their head dynamics, and it can effectively adapt to various riding conditions and individuals. The system incorporates a deep learning framework with an advanced domain adversarial network. By single-time pre-training, HeadMon+ is capable of adapting to new data domains, including human subjects, and riding conditions for robust maneuver prediction. Based on our evaluation, we have found that the maneuver prediction of HeadMon+ has an overall precision of 94% with a prediction time gap of 4 seconds. HeadMon+'s low cost and rapid response capability make it easily deployed and then contribute to enhancing safe riding.

This study investigates the bidirectional relationship between emotional changes and physical activity, specifically focusing on the number of steps taken. Using Propensity Score Matching (PSM), we analyzed how fluctuations in emotional states influence physical activity and, conversely, how increases/decreases in daily steps impact subsequent emotional well-being. This study uses a dataset containing data on daily steps and emotional status collected via a smartphone application (N=123). Our findings indicate that increases in the number of steps significantly increase the subsequent emotion report positively. Additionally, changes in emotional status have relations with a subsequent number of steps. These results suggest a reciprocal influence between emotional and physical activities, highlighting the importance of integrating physical and mental health interventions.

There will be an increase in situations where images taken in specific locations are transmitted through networks for various services. However, this trend can lead to significant communication loads due to simultaneous transmission of images from multiple locations. Therefore, it is important to reduce the amount of network traffic in image transmission. While traditional compression methods focus on minimizing information loss in images, some applications only require the retention of semantic information, suggesting potential improvements in communication efficiency. This paper proposes an image generation-based transmission method for highly-efficient communications exploiting composition and descriptive information. The proposed method extracts specific information from an image to decrease the amount of data transmission, and reconstructs the image using an image-generative model by a receiver. In addition, image compression and reconstruction in the proposed method are demonstrated through an experiment. The experimental results indicate a need for a method to evaluate the output and a method for image reconstruction based on this evaluation.

We propose a user mobility-driven federated learning method, which integrates learning models from different regions, leveraging user mobility. This method aims to improve performance of learning models in specific regions by merging them with models from other areas. In regions with less user mobility, our method creates unique regional models, while in areas with high mobility, it integrates models for enhanced performance. Evaluation results indicate that accuracy improved with additional training, although it temporarily decreased after model integration.

Silent speech interaction (SSI) allows users to discreetly input text without using their hands. Existing wearable SSI systems typically require custom devices and are limited to a small lexicon, limiting their utility to a small set of command words. This work proposes ReHEarSSE, an earbud-based ultrasonic SSI system capable of generalizing to words that do not appear in its training dataset, providing support for nearly an entire dictionary’s worth of words. As a user silently spells words, ReHEarSSE uses autoregressive features to identify subtle changes in ear canal shape. ReHEarSSE infers words using a deep learning model trained to optimize connectionist temporal classification (CTC) loss with an intermediate embedding that accounts for different letters and transitions between them. We find that ReHEarSSE recognizes 100 unseen words with an accuracy of 89.3%.

As the capabilities of smart sensing and mobile technologies continue to evolve and expand, storing diverse sensor data on smartphones and cloud servers becomes increasingly challenging. Effective data compression is crucial to alleviate these storage pressures. Compressed sensing (CS) offers a promising approach, but traditional CS methods often struggle with the unique characteristics of sensor data-like variability, dynamic changes, and different sampling rates-leading to slow processing and poor reconstruction quality. To address these issues, we developed Mob-ISTA-1DNet, an innovative CS framework that integrates deep learning with the iterative shrinkage-thresholding algorithm (ISTA) to adaptively compress and reconstruct smartphone sensor data. This framework is designed to manage the complexities of smartphone sensor data, ensuring high-quality reconstruction across diverse conditions. We developed a mobile application to collect data from 30 volunteers over one month, including accelerometer, gyroscope, barometer, and other sensor measurements. Comparative analysis reveals that Mob-ISTA-1DNet not only enhances reconstruction accuracy but also significantly reduces processing time, consistently outperforming other methods in various scenarios.

Child-rearing depression, triggered by the chronic stress of parenting, can lead to serious mental health issues if not detected early. This study uses passive mobile sensing to analyze the behavioral and psychological patterns of households with preschool children. By collecting data from 131 participants (including 18 parents of preschoolers), we aim to differentiate child-rearing anxiety and behavior patterns. Our focus includes step counts, location data, call frequency, and psychological states. Results indicate that parents of preschoolers have fewer steps, visit fewer locations, and have higher call activity. They also show higher stress and anxiety but lower depression levels, suggesting that family support may mitigate depressive symptoms. These insights could aid in developing early detection and intervention strategies for child-rearing depression.

Although micro-mobility has become a popular and indispensable mode of transportation in recent years, it has also introduced a large number of traffic accidents. Timely tracking and predicting the maneuvers hold the potential to prevent accidents through prompt warnings and interventions. However, the open and simple structure of micro-mobility makes it hard to install sophisticated infrastructures for maneuver prediction. In this paper, we argue that the micro-mobility body dynamics provide sufficient information for maneuver prediction. Our preliminary study suggests that micro-mobility body dynamic patterns appear beforehand and exhibit the correlation with steering maneuvers. We accordingly present RideGuard, which leverages a built-in Inertial Measurement Unit on smartphones to achieve the prediction of steering maneuvers. Through a dual-stream CNN deep learning architecture, RideGuard effectively captures complex patterns and feature relationships from the time and frequency domain. Our extensive real-traffic experiments involving 20 participants demonstrate the superiority of RideGuard: employing a 3s detection window, RideGuard attains a minimum of 94% precision in maneuver prediction with a 5s prediction time gap. The low-cost and rapid response feature of RideGuard enables feasible deployment and promotes safer riding practices. Additionally, we open-source our well-labeled dataset to facilitate further research.

Monitoring and controlling the exposure of an individual to ultraviolet (UV) radiation is crucial for personal health. The use of the global navigation satellite system (GNSS) signals received by a personal off-the-shelf smartphone has been studied as a novel estimation method. In the existing method, satellites are grouped based on their positions and the signal information is represented by group statistics, leading to a coarse estimation. We propose a new UV index estimation method that directly utilizes satellite-wise information and their spatial relationships with a point-cloud neural network, considering the similarity between GNSS signals and point clouds. We collected GNSS signals and UV index data from two locations within the same area and demonstrated that the proposed method enhances the estimation accuracy and smoothness.

Public awareness of automated external defibrillator (AED) locations is crucial for prompt retrieval in cardiac emergencies. We propose a simulation-based approach as a preliminary step towards developing gamified mobile apps to enhance this awareness. By simulating AED retrieval in real-world pedestrian networks under various scenarios, we identify key elements that can improve retrieval efficiency. Our findings confirm the viability of the framework and highlight crucial aspects for improvement towards efficient future applications.

Ahstract-Various methods focusing on either spatial or temporal utilization, such as path diversity or link utilization, have been studied to realize efficient ad hoc networks. Furthermore, a method that integrates these methods has been proposed; however, it has only been evaluated theoretically and partially. Hence, achieving efficiency from both viewpoints as a protocol is challenging. This study proposes and defines block transmission-enabled opportunistic routing, which can achieve spatiotemporal efficiency as a protocol. The proposed method was evaluated by comparing its performance with that of two conventional methods that focus on either spatial or temporal efficiency, through computer simulations.

Machine learning (ML) is predominantly performed in the cloud or other computing facilities. While this computing method allows for the benefits of ML to be leveraged in urban settings, it may also lead to unfair sharing of the environment owing to the heat generated by servers in areas housing computing infrastructure. In this study, we developed a green distributed machine leaning (DML) concept-CoopL-to calculate the local consumption of computational resources based on DML, thereby mitigating the environmental burden. Moreover, we analyzed the impact of long-distance communication by incrementally raising the communication latency. Furthermore, the power consumption was examined by considering the hop count of the router.

Federated learning (FL) achieves collaborative learning without the need for data sharing, thus preventing privacy leakage. To extend FL into a fully decentralized algorithm, researchers have applied distributed optimization algorithms to FL by considering machine learning (ML) tasks as parameter optimization problems. Conversely, the consensus-based multi-hop federated distillation (CMFD) proposed in the authors' previous work makes neural network (NN) models get close with others in a function space rather than in a parameter space. Hence, this study solves two unresolved challenges of CMFD: (1) communication cost reduction and (2) visualization of model convergence. Based on a proposed dynamic communication cost reduction method (DCCR), the amount of data transferred in a network is reduced; however, with a slight degradation in the prediction accuracy. In addition, a technique for visualizing the distance between the NN models in a function space is also proposed. The technique applies a dimensionality reduction technique by approximating infinite-dimensional functions as numerical vectors to visualize the trajectory of how the models change by the distributed learning algorithm.

Although micro-mobility brings convenience to modern cities, they also cause various social problems, such as traffic accidents, casualties, and substantial economic losses. Wearing protective equipment has become the primary recommendation for safe riding. However, passive protection cannot prevent the occurrence of accidents. Thus, timely predicting the rider's maneuver is essential for active protection and providing more time to avoid potential accidents from happening. Through the qualitative study, we argue that we can use the rider's head dynamic as an information source to predict the rider's following maneuvers. We accordingly present HeadMon, a riding maneuver prediction system for safe riding. HeadMon utilizes the head dynamics of a rider by installing an inertial measurement unit on the helmet. It uses the extracted head dynamics features as the input of the deep learning architecture to achieve prediction. We implemented the HeadMon prototype on Android smartphone as a proof of concept. Through comprehensive experiments with 20 participants, the result demonstrates the excellent performance of HeadMon: not only could it achieve an overall precision of at least 85% for maneuver prediction under a 4s prediction time gap, but it also could keep a high accuracy under a low sampling rate. The low-cost feature of HeadMon allows it to be readily deployable and towards more safety riding.

Distracted riding behavior is one of the main causes of bicycle-related traffic accidents, resulting in a large number of casualties and economic losses every year. There is an urgent need to address this problem by accurately detecting distracted riding behaviors. Inspired by the observation that distracted riding behaviors induce unique head motion features that respond to the rider’s attention, we present the HeadSense, a helmet-based system that not only monitors the visual search episode of the rider but also detects distracted riding behaviors. Specifically, HeadSense leverages the inertial motion unit (IMU) to recognize distracted behaviors such as using smartphones, attracting to the roadside element, and abreast riding. We designed, implemented, and evaluated HeadSense through extensive experiments. We conducted experiments with 19 participants inside the university’s campus. The experimental results show that HeadSense can achieve an overall accuracy of 86.14% while monitoring visual search episodes. Moreover, HeadSense can detect the occurrence of distracted riding behaviors with an average precision of up to 85.04%.

In this letter, we investigate the problem of detecting compressed stochastic sparse signals with unknown sparsity degree under Bernoulli–Gaussian model. In addition to the generalized likelihood ratio test (GLRT) proposed in [1], the corresponding Rao test and Wald test are derived in this letter. By observing that obtaining their analytical performance is challenging, we further propose a new probability constraint estimator (PCE) of the unknown sparsity degree. Interestingly, by adopting the PCE, the GLRT, Rao and Wald tests are shown to be statistically equivalent and reduce to a new detector (i.e., the detector with PCE) with a simple structure. The analytical performance of the detector with PCE is thus derived, which is verified by Monte Carlo simulations. Finally, numerical experiments illustrate that the proposed Rao test and the detector with PCE outperform the original GLRT.

Trip purpose plays a critical role in reflecting human mobility behavior. However, it is relatively difficult to determine. With the rapid growth of urban mobility and big mobile data, utilizing these data for trip purpose classification has been a long-term objective to enhance travel demand and behavior models used in urban planning. Although studies on this topic have been extensively conducted, most past research preferred relying on traveler attributes or long-term travel histories to achieve accurate results. These data could be privacy sensitive and often do not satisfy real-world scenarios. This study addresses the problem of classifying trip purpose by only space activity information to avoid privacy conflict. 1) External memories are collected from factorized components based on the non-negative Tucker decomposition scheme. 2) These memories are extended by the cross-attention mechanism to achieve feature augmentation. 3) Subsequently, a novel concept called "latent mode alignment" is proposed. By leveraging the linear characteristics of external memories, geographic contextual latent modes are represented and matched with travel activities; this procedure is called "alignment." 4) The gate mechanism controls the eventual outputs for update. The proposed plug-in memory network (PMN), combined with baseline models, effectively outperforms the original settings. Moreover, combination models are validated with strong tolerance through missing data tests, which are common and problematic in real-world scenarios. The proposed PMN is a plug-and-play design that is easy to combine with newly developed classification models, and other memory collection methods can be expected.

Wearing a mask considerably mitigates the risk of infection from droplets. Automatic detection of whether a person wears a mask in his/her daily life and the type of masks the person wears can provide useful information for various services such as infection risk assessment, just-in-time alerts, and lifelogging. However, such automatic detection is difficult without the use of video processing or specialized equipment. In this study, the motion sensor of a commercially available smartwatch was used to detect the mask-wearing status. An investigation of the acceleration characteristic and an evaluation experiment of the mask-wearing state detection model revealed an accuracy of approximately 90% when specific motions were classified using motion sensors and machine learning. Furthermore, 98% accuracy was achieved when classifying sitting and walking activities.

Users and operators of shared spaces must ensure safety in such areas to prevent the spread of COVID-19. Although each organization has operated a variety of safety-related systems, including contact tracing, congestion monitoring, and check-in services, it is unclear what elements, such as privacy protection level, benefits, and permission procedures, have promoted the usage of these systems. In this study, we created MOCHA, a platform for sharing and tracking room-level locations. This platform automatically detects visited places by scanning Bluetooth beacons in each room using smartphones and shares location data according to predefined user settings. The collected data is used for room-level contact tracing, congestion monitoring, and reservation services. According to >6,500 users' usage data for a year in a university, outlining the advantages of utilizing the app encouraged people to install the app, and reinforced connections in small private groups are encouraged to use the app continuously.

While micro-mobility brings convenience to the modern city, they also cause various social problems such as traffic accidents, casualties, and huge economic losses. Wearing protective equipment has become the primary recommendation for safe riding, but passive protection cannot prevent accidents from happening after all. Thus, timely predicting the rider's maneuver is essential for more active protection and buying more time to avoid potential accidents from happening. In this poster, we explore the feasibility of using riders' head dynamics to predict their riding maneuvers. Through ten participants' preliminary study, we observed that not only do riders' head movements appear ahead of their maneuvers but also head movement patterns are distinct with different maneuver intentions. We then construct a deep learning network using Long Short Term Memory, achieving 89% of accuracy on maneuver prediction.

Monitoring the amount of UV irradiance to which individuals are exposed and ensuring that every individual receives the optimal amount has been the subject of extensive research. In previous research, the UV index was estimated using cell phone cameras, light sensors on smartphones, or wearable UV sensors. We propose a method for estimating the UV index using the widespread global navigation satellite system (GNSS) sensors available on smart-phones. In contrast to approaches that require the sensor to be exposed continuously to the irradiance, this method, which leverages GNSS sensors, has the potential advantage of enabling UV index measurement simply by carrying the phone as usual. As a first step in measuring the index using GNSS sensors, GNSS data were collected from cell phones placed at three locations in a single area; the OpenUV API was utilized as a baseline. The proposed method achieved a mean absolute error of 0.1523, which significantly outperformed the baseline.

When customers wait their turn to place an order at a street vendor, they often form a spontaneous queue. Due to the presence of passers-by and people standing outside the queue, it is more difficult than one might think to distinguish between those in the queue and those not in the queue. In this paper, we consider a method that uses acoustic ranging to autonomously detect who is in line and in which order, under the condition that all customers have smartphones. The proposed method is unique in that it can distinguish whether a newly arrived user has joined the end of the queue or not by taking cues from the geometric properties of the queue. Our preparatory queueing simulations confirm that 92.5% of the queuers are estimated correctly.

Indoor localization technology is one of the most important topics in the fields of Internet of Things (IoT) and ubiquitous computing, and it has attracted much attention in recent years due to its ability to enable a variety of services. Localization methods using Bluetooth or Wi-Fi signal strength can introduce a low-cost location estimation system. However, due to the instability of the received signal strength and signals leaking from adjacent rooms, a simple method based on signal strength alone frequently results in misjudgment, depending on the signal propagation characteristics. In this paper, we propose a method to suppress misjudgment by considering the characteristics of stay time in different rooms. Our proposed method estimates the user state by fitting the distribution of time spent in each room to a Weibull distribution and applying survival analysis. The experimental results suggest that the method will provide more accurate information about the rooms in which users stay.

As intelligent sensing and smartphone technologies have progressed, a huge amount of highly heterogeneous data have come to be stored in smartphones and uploaded to servers for analysis on a daily basis. This has led to vast storage overheads for users and companies. Hence, data compression becomes the most efficient strategy for suppressing the increase in storage overhead. Compressed sensing (CS) technology is one approach to compressing data, but traditional CS-based algorithms are significantly time-consuming and have low reconstruction performance. In light of these drawbacks, this paper proposes a compressed sensing framework that instead takes advantage of the low time cost and adaptive learning capability of deep learning methods, wherein a convolutional neural network (CNN) is used for compressing and reconstructing acceleration data. Our experiments with actual smartphone acceleration data show that the proposed method dramatically improves the reconstruction performance with very little reconstruction time compared with traditional compressed sensing methods.

The childcare environment has significantly changed, owing to accelerating women's social advancement and the increasing number of nuclear families. Improving and supporting childcare have become major challenges in current society. Automatically recording and subsequently observing childcare activities can be used for various purposes to support childcare. However, methods to detect childcare activities using off-the-shelf devices have not yet been proposed. This study develops a method to detect childcare activities that parents perform for their babies using an off-the-shelf wearable device. We define nine childcare activities and develop corresponding detection models based on motion-sensor data from a smartwatch. Our evaluation in a laboratory setting resulted in classification performances of 71% (F1: 0.66).

The frequency and duration of social contact, represented by conversation, is positively correlated with our physical and mental health. Therefore, a method that automatically measures social contact can provide insight into people's health conditions and risks. Even though off-the-shelf wristwatch-type wearable devices are widely used in our daily lives and have rich computational resources, they have not been used as a social-contact monitoring tool in everyday conditions. In this study, we propose a system, called Ohanashi, for continuously monitoring conversational events as a means of social contact in daily life, by edge processing on an off-the-shelf smartwatch. To monitor the conversational event, we developed an audio classification model and implemented it as a smartwatch application, which can classify conversation and noise from an audio stream. Our performance evaluation shows that the classification model can classify conversation and noise with more than 86% accuracy in both silent and noisy environments, and the system can monitor conversation events for more than 15 hours on a smartwatch.

Riding bicycles with only one hand on the handlebar can severely undermine the operator’s steering capability and threaten road and transportation safety. Prior studies have exploited motion sensors to detect riding contexts and recognize related behaviors. Nevertheless, they fail to integrate a scheme to account for single-handed riding with elements crucial to danger prevention: awareness of the surroundings, response to danger, and convenient adoption. In this work, we proposed, designed, and implemented DoubleCheck: a smartphone-based real-time framework for cycling hand detection and distraction recognition. The method monitors handlebar holding on different road surfaces and recognizes hazardous distraction activities related to single-handed cycling using motion signals captured by a built-in inertial measurement unit in a handlebar-borne smartphone. It was designed on the premise that single-handed cycling enabled operators to adapt their body movements to different (often distracting) activities. We conducted an evaluation experiment using 22 participants on asphalt and pavement. The results indicate that DoubleCheck achieves an F1-score of 0.96 for hand detection and 0.69 for distraction recognition, demonstrating its efficacy as a candidate rider-safety precautionary measure.

Owing to lifestyle changes, urbanization, and the COVID-19 pandemic, many people spend more time indoors and tend to receive less direct sunlight than before. Although excessive or inadequate ultraviolet (UV) exposure can be harmful to our physical and mental health, moderate UV exposure is essential for vitamin D (VD) production in the body. In this study, we estimate the UV exposure using an off-the-shelf smartphone, and explore the relationship between the UV values and GPS signal strength (C/N0). The results demonstrate that a strong correlation (R2= 0.73) between the UV values and carrier to noise density (C/N0) even if the smartphone and UV sensor are moved. Therefore, it is possible to estimate the UV exposure to some extent from a person's location, even while carrying a smartphone.

In vehicular edge computing (VEC), offloading the tasks to the nearby resource-rich edge servers helps each vehicle enhance computational capabilities and improve in-vehicle applications' performance. However, the concentration of travel at specific spaces and times poses significant challenges on the load-balancing and scheduling of computation tasks at the edge servers. This paper studies a low-complexity dynamic online offloading strategy that efficiently reduces task delay and computing resource consumption in the multi-user, multiserver vehicular edge computing scenarios. Our design addresses issues of computation task placement and execution order of the tasks on each server. We use a realistic approach that vehicles generate tasks over time, and the set of the tasks is unknown in advance so that the offloading decisions are made in runtime. Extensive simulations are conducted on a real mobility trace of Luxembourg city, and the results show that the proposed algorithm effectively improves the offloading utility of the system.

Contagious diseases like COVID-19 spread periodically and threaten our lives. Self-protection, such as washing hands, wearing a mask, and staying home, are simple and practical solutions to safeguard against these diseases. Most governments and health departments recommend that people maintain self-protection. Although continuous self-protection effectively prevents the spread of infection, only the intent to self-protect is unsustainable in the long term. In this study, we design, develop, and deploy an application to track users' daily activities semi-automatically and enhance self-protection behavior using mobile sensing and gamified feedback techniques. Currently, more than 324 people have installed the app via AppStore, and 52 users have shared their activity data to our research group.

Parkinson's disease (PD) is the second most common neurodegenerative disorder, impacting an estimated seven to ten million people worldwide. Measuring the symptoms and progress of the disease, and medication effectiveness is currently performed using subjective measures and visual estimation. We developed and evaluated a mobile application, STOP for tracking hand's motor symptoms, and a medication journal for recording medication intake. We followed 13 PD patients from two countries for a 1-month long real-world deployment. We found that PD patients are willing to use digital tools, such as STOP, to track their medication intake and symptoms, and are also willing to share such data with their caregivers and medical personnel to improve their own care.

We propose an integrated framework for sensing, recognizing and utilizing of subjective time as context. Various studies on experimental psychology have showed several factors which affects subjective time. Those factors should be partially captured by ubiquitous sensors such as smartphones and wearable devices, therefore, we tackle to create common and individual model for subjective time based on the sensor data. We report our first prototype implementation for the framework based on AWARE framework with adding experience sampling method for subjective time recognition. In addition, we discuss potential applications which leveraging advantages of subjective time as context.

Parkinson's disease (PD) is the second most common neurodegenerative disorder, impacting an estimated seven to ten million people worldwide. It is commonly accepted that improving medication adherence alleviates symptoms and maintains motor capabilities. Not following the medication regimen (e.g., skipping or over-medicating) may worsen side-effects, which mislead clinicians and patients. We developed and evaluated a mobile application, STOP, for screening the PD symptoms and medication intake. It contains a game for tracking the PD symptoms, and a medication journal for recording medical intake and adherence. We conducted a 1-month long real-world deployment with 13 PD patients from two countries. We found that the application medication adherence tracking provides non-bias information, and users are receptive to share such data with their care and medical personnel.

We propose "SmileWave", the first selfie social networking service to reveal the existence of emotional cognation through smiling selfies on the social network. We conducted multiple rounds of in-the-wild user studies with 86 cumulative total users for total duration of 5 weeks. Throughout the entire study, we confirmed the occurrence of smile-based emotional contagion over social network, not only in the momentary duration but in longer term period.

This paper proposes ''SmileWave", a system for revealing smile-based emotional contagion, propagation effect of the similar emotion through smiley facial expression, on the social network where users interact each other through web-based user interface rather than in-person interaction. SmileWave is a picture-based networking service and detects the change of smile degree when the user looks at posted smile images of others. Our extensive user study with 50 participants for 30 days confirmed the emotional contagion effect on SmileWave. Users' smile degree improved by 27% when the user looked at posted smile images. The result also proved that there is a stronger effect on smile-based emotional contagion when the examinee and the person in the image are in close relationship.

Recent technological trends in mobile/wearable devices and sensors have been enabling an increasing number of people to collect and store their "lifelog" easily in their daily lives. Beyond exercise behavior change of individual users, our research focus is on the behavior change of teams, based on lifelogging technologies and lifelog sharing. In this paper, we propose and evaluate six different types of lifelog sharing models among team members for their exercise promotion, leveraging the concepts of "competition" and "collaboration." According to our experimental mobile web application for exercise promotion and an extensive user study conducted with a total of 64 participants over a period of three weeks, the model with a "competition" technique resulted in the most effective performance for competitive teams, such as sports teams.

Recent technological trends on mobile/wearable devices and sensors have been enabling increasing number of people to collect and store their “life-logs” easily in their daily lives. Beyond exercise behavior change of individual user, our research focus is on the behavior change of teams, based on life-logging technologies and information sharing. In this paper, we propose and evaluate six different types of information sharing model among team members for their exercise promotion, leveraging concepts of “competition” and “collaboration”. According to our experimental mobile web application for exercise promotion and extensive user study among 64 total users for three weeks, the model with “external competition” technique resulted the most effective performance for competitive teams such as sport teams.