This paper presents a positioning system for "people-centric" navigation, which estimates relative positions of surrounding people to help users to find a target person in a crowd of neighbors. Our system, called PCN, employs pedestrian dead reckoning (PDR) and proximity sensing with Bluetooth only using off-the-shelf mobile phones. Utilizing the feature of "group activity" where people naturally form groups moving similarly and together in exhibitions, parties and so on, PCN corrects deviation of distance and direction in PDR. The group information is also helpful to identify the surrounding people in the navigation. A field experiment in a real exhibition with 20 examinees carrying Google Android phones was conducted to show its effectiveness.
The behavior observations on young children in new, first-in-the-life environments have significant implications. We can often uniquely observe a child's unforeseen interaction with the environment and peer-children. It would be not only a piece of discovery but a beginning of an open quest worth exploring. Out-of-classroom activities like kindergarten's field trips are perfect opportunities, but those are quite different from regular classroom activities where the teachers' conventional observation methods are hardly practical. This paper proposes a novel approach to extend the teachers' awareness on the children's field trip behaviors by means of mobile and sensor technology. We adopt the notion of behavioral distribution and singularities. We estimate the children's representative behavioral state in a given context, and study the effect of focusing on the behaviors which are unlikely in this context. We discuss our 14-month collaborative study and various qualitative benefits through multiple deployments on actual kindergarten field trips.
Understanding the effect of pervasive services on user context is critical to many context-aware applications. Detailed descriptions of context-altering services are necessary, and manually adapting them to the local environment is a tedious and error-prone process. We present a method for automatically providing service descriptions by observing and learning from the behavior of a service with respect to its environment. By applying machine learning techniques on the observed behavior, our algorithms produce high quality localized service descriptions. In a series of experiments we show that our approach, which can be easily plugged into existing architectures, facilitates context-awareness without the need for manually added service descriptions.
Commercial home automation systems are becoming increasingly common, affording the opportunity to study technology-augmented homes in real world contexts. In order to understand how these technologies are being integrated into homes and their effects on inhabitants, we conducted a qualitative study involving smart home professionals who provide such technology, people currently in the process of planning or building smart homes, and people currently living in smart homes. We identified motivations for bringing smart technology into homes, and the phases involved in making a home smart. We also explored the varied roles of the smart home inhabitants that emerged during these phases, and several of the challenges and benefits that arise while living in a smart home. Based on these findings we propose open areas and new directions for smart home research.
Emerging electric-drive vehicles, such as hybrid electric vehicles (HEVs) and plug-in HEVs (PHEVs), hold the potential for substantial reduction of fuel consumption and greenhouse gas emissions. User driving behavior, which varies from person to person, can significantly affect (P)HEV operation and the corresponding energy and environmental impacts. Although some studies exist that investigate vehicle performance under different driving behaviors, either directed by vehicle manufacturers or via on-board diagnostic (OBD) devices, they are typically vehicle-specific and require extra device/effort. Moreover, there is no or very limited feedback to an individual driver regarding how his/her personalized driving behavior affects (P)HEV performance. This paper presents a personalized driving behavior monitoring and analysis system for emerging hybrid vehicles. Our design is fully automated and non-intrusive. We propose phone-based multi-modality sensing that captures precise driver–vehicle information through de-noise, calibration, synchronization, and disorientation compensation. We also provide quantitative driver-specific (P)HEV analysis through operation mode classification, energy use and fuel use modeling. The proposed system has been deployed and evaluated with real-world user studies. System evaluation demonstrates highly-accurate (0.88-0.996 correlation and low error) driving behavior sensing, mode classification, energy use and fuel use modeling.
Researchers in pervasive and ubiquitous computing have produced much work on new sensing technologies for disaggregating domestic resource consumption, and on designs for energy-centric interventions at home. In a departure from this, we employ a service-oriented approach, where we account for not only the amount of resources that specific appliances draw upon, but also how the associated services may be characterised in the context of everyday life. We undertook a formative study in four student flats over a twenty-day period, collecting data using interviews with eleven participants and over two hundred in-home sensors. Following an in-depth description of observations and findings from our study, we argue that our approach provides a more inclusive range of understandings of resources and everyday life than has been shown from energy-centric approaches.
This paper describes the development of a new finger-ring shaped sensor device with a coil of wire for recognizing the use of handheld electrical devices such as digital cameras, cellphones, electric toothbrushes, and hair dryers by sensing time-varying magnetic fields emitted by the devices. Recently, sensing the usage of home electrical devices has emerged as a promising area for activity recognition studies because we can estimate high-level daily activities by recognizing the use of electrical devices that exist ubiquitously in our daily lives. A feature of our approach is that we can recognize the use of electrical devices that are not connected to the home infrastructure without the need to equip them with sensors. We evaluated the performance of our approach by using sensor data obtained from real houses. We also investigated the portability of training data between different users.
Living in society, to go out is almost inevitable for healthy life. There is increasing attention to it in many fields, including pervasive computing, medical science, etc. There are various factors affecting the daily going-out behavior such as the day of the week, the condition of one's health, and weather. We assume that a person has one's own rhythm or patterns of going out as a result of the factors. In this paper, we propose a non-parametric clustering method to extract one's rhythm of the daily going-out behavior and a prediction method of one's future presence using the extracted models. We collect time histories of going out/coming home (6 subjects, total 827 days). Experimental results show that our method copes with the complexity of patterns and flexibly adapts to unknown observation.
In this paper we propose and implement a battery-shaped sensor node that can monitor the use of an electrical device into which it is inserted by sensing the electrical current passing through the device. We live surrounded by large numbers of electrical devices and frequently use them in our daily lives, and so we can estimate high-level daily activities by recognizing their use. Therefore, many ubiquitous and wearable sensing studies have attempted to recognize the use of electrical devices by attaching sensor nodes to the devices directly or by attaching multiple sensors to a user. With our node, we can easily monitor the use of an electrical device simply by inserting the node into the battery case of the device. We also propose a method that automatically identifies into which electrical device the sensor node is inserted and recognizes electrical events related to the device by analyzing the current sensor data. We evaluated our method by using sensor data obtained from three real houses and achieved very high identification and recognition accuracies.
Wireless network operators increasingly deploy WiFi hotspots and low-power, low-range base stations in order to satisfy users' growing demands for context-aware services and performance. In addition to providing better service, such capillary infrastructure deployment threatens users' privacy with respect to their social ties and communities, as it allows infrastructure owners to infer users' daily social encounters with increasing accuracy, much to the detriment of their privacy. Yet, to date, there are no evaluations of the privacy of communities in pervasive wireless networks. In this paper, we address the important issue of privacy in pervasive communities by experimentally evaluating the accuracy of an adversary-owned set of wireless sniffing stations in reconstructing the communities of mobile users. During a four-month trial, 80 participants carried mobile devices and were eavesdropped on by an adversarial wireless mesh network on a university campus. To the best of our knowledge, this is the first study that focuses on the privacy of communities in a deployed pervasive network and provides important empirical evidence on the accuracy and feasibility of community tracking in such networks.
Differences in perceptual and cognitive abilities between the young and elderly have implications for in-car tasks. As a primary example, although in-car navigation systems enhance situational awareness, this comes at the cost of increasing visual distraction and cognitive load. To address these shortcomings, this paper explores the efficacy of multi-modal cues for providing route guidance information. We present the results of a study evaluating the impact of multi-modal feedback on driving performance and cognitive load. We found that the full combination of visual, auditory, and haptic feedback was generally most useful to reduce way-finding errors. However, our study highlighted a number of differences between elder and younger drivers for their safer navigation. Adding more modalities strained the already high workload of elder drivers. In contrast, adding haptic feedback to traditional audio and visual feedback led to more attentive driving by younger drivers. Therefore, for elder drivers, navigation systems need to be personalized to enhance the benefit of auditory feedback without increasing the number of sensory feedbacks. For younger drivers, it is necessary to incorporate new non-visual feedback to minimize distractions caused by visual feedback. We demonstrate these results through task performance-based measures, subjective workload measures and through objective workload measures that use psychophysiological responses of participants to predict a driver’s cognitive load in near real-time.
RFID technology is now widely used to identify, locate, track and monitor physical objects. However, the use of RFID technology in modern manufacturing has been limited because of the unreliability of RFID devices. In addition to this, where it is used, the technology is mostly deployed to be a substitute for manual inventory management. In this paper we present the Smart Factory, a modern factory infrastructure capable of monitoring each and every product part that moves across the factory during the entire production process. In order to overcome the reliability issues in RFID devices, we have built up a probabilistic model to assign probabilities to the RFID readers and to the product part detections. We also present a probability self-calibration algorithm that automatically adapts the probabilities of RFID readers to better reflect their performance at current instance of time.
The increasing popularity of location based social services such as Facebook Places, Foursquare and Google Latitude, solicits a new trend in fusing social networking with real-world sensing. The availability of a wide range of sensing technologies in our everyday environment presents an opportunity to further enrich social networking systems with fine-grained real-world sensing. However, the introduction of passive sensing into a social networking application disrupts the traditional, user-initiated input to social services, raising both privacy and acceptability concerns. In this work we present an empirical study of the introduction of a sensor-driven social sharing application within the working environment of a research institution. Our study is based on a real deployment of a system that involves location tracking, conversation monitoring, and interaction with physical objects. By utilizing surveys, interviews and experience sampling techniques, we report on our findings regarding privacy and user experience issues, and significant factors that can affect acceptability of such services by the users. Our results suggest that such systems deliver significant value in the form of self reflection and comparison with others, while privacy concerns are raised primarily by the limited control over the way individuals are projected to their peers.
WiFi fingerprinting is currently one of the most popular techniques for indoor localization as it provides reasonable positioning accuracy while at the same time being able to exploit existing wireless infrastructure. To facilitate calibration efforts and to overcome fluctuations in location measurements, many indoor WiFi positioning systems utilize a discrete partitioning, e.g., a grid or a topological map, of the space where the positioning is being deployed. A major limitation of this approach, however, is that instead of considering spatial similarities in the signal environment, the partitioning is typically based on an uniform division of the space or topological constraints (e.g., rooms and walls). This can significantly decrease positioning accuracy when the signal environment is not sufficiently stable across all partitions. Moreover, current solutions provide no support for identifying partitions that are not compatible with the current wireless deployment. To overcome these limitations, we propose AWESOM (Activations Weighted by the Euclidean-distance using Self-Organizing Maps), a novel measure for automatically creating a discrete partitioning of the space where the WiFi positioning is being deployed. In addition to enabling automatic construction of a discrete partitioning, AWESOM provides a measure for evaluating the goodness of a given partitioning for a particular access point deployment. AWESOM also enables identifying partitions where additional access points should be deployed. We demonstrate the usefulness of AWESOM using data collected from two large scale deployments of a proprietary wireless positioning system in a hypermarket environment.
This paper presents two novel handheld projector systems for indoor pervasive computing spaces. These projection-based devices are "aware" of their environment in ways not demonstrated previously. They offer both spatial awareness, where the system infers location and orientation of the device in 3D space, and geometry awareness, where the system constructs the 3D structure of the world around it, which can encompass the user as well as other physical objects, such as furniture and walls. Previous work in this area has predominantly focused on infrastructure-based spatial-aware handheld projection and interaction. Our prototypes offer greater levels of environment awareness, but achieve this using two opposing approaches; the first infrastructure-based and the other infrastructure-less sensing. We highlight a series of interactions including direct touch, as well as in-air gestures, which leverage the shadow of the user for interaction. We describe the technical challenges in realizing these novel systems; and compare them directly by quantifying their location tracking and input sensing capabilities.
A Segway is often used to transport a user across mid range distances in urban environments. It has more degrees of freedom than car/bike and is faster than pedestrian. However a navigation system designed for it has not been researched. The existing navigation systems are adapted for car drivers or pedestrians. Using such systems on the Segway can increase the driver's cognitive workload and generate safety risks. In this paper, we present a Segway AR-Tactile navigation system, in which we visualize the route through an Augmented Reality interface displayed by a mobile phone. The turning instructions are presented to the driver via vibro-tactile actuators attached to the handlebar. Multiple vibro-tactile patterns provide navigation instructions. We evaluate the system in real traffic and an artificial environment. Our results show the AR interface reduces users' subjective workload significantly. The vibro-tactile patterns can be perceived correctly and greatly improve the driving performance.
Retailers and brands are just starting to utilize online social media to support their businesses. Simultaneously, public displays are becoming ubiquitous in public places, raising the question about how these two technologies could be used together to attract new and existing customers as well as strengthen the relationship toward a focal brand. Accordingly, in a field experiment we displayed brand- and product-related comments from the social network Facebook as pervasive advertising in small-space retail stores, known as kiosks. From interviews conducted with real customers during the experiment and the corresponding analysis of sales data we could conclude three findings. Showing social media comments resulted in (1) customers perceiving brands as more innovative and attractive, (2) a measurable, positive effect on sales on both the brand and the product in question and (3) customers wanting to see the comments of others, but not their own, creating a give-and-take paradox for using public displays to show social media comments.
A probabilistic method for inferring common routes from mobile communication network traffic data is presented. Besides providing mobility information, valuable in a multitude of application areas, the method has the dual purpose of enabling efficient coarse-graining as well as anonymisation by mapping individual sequences onto common routes. The approach is to represent spatial trajectories by Cell ID sequences that are grouped into routes using locality-sensitive hashing and graph clustering. The method is demonstrated to be scalable, and to accurately group sequences using an evaluation set of GPS tagged data.
In developing regions, the reach of crowdsourcing services such as Amazon Mechanical Turk (mTurk) has been limited by the lack of adequate payment mechanisms and low visibility amongst the crowd. In this paper, we present a commodity based model for crowdsourcing where crowd workers get paid in kind in the form of a commodity instead of money. Our model makes crowdsourcing services more visible to users in developing regions and also addresses the issue of payment. We conducted two field studies in urban India to evaluate the applicability of our proposed model. Our results show that the commodity based crowdsourcing model reached workers with very different demographics from the typical mTurk workers. We also found that users preferred to receive a commodity instead of money as remuneration.
.NET Gadgeteer is a new platform conceived to make it easier to design and build custom electronic devices and systems for a range of ubiquitous and mobile computing scenarios. It consists of three main elements: solder-less modular electronic hardware; object-oriented managed software libraries accessed using a high-level programming language and established development environment; and 3D design and construction tools designed to facilitate a great deal of control over the form factor of the resulting electronic devices. Each of these elements is designed to be accessible to a wide range of people with varying backgrounds and levels of experience and at the same time provide enough flexibility to allow experts to build relatively sophisticated devices and complex systems in less time than they are used to. In this paper we describe the .NET Gadgeteer system in detail for the first time, explaining a number of key design decisions and reporting on its use by new users and experts alike.
