Philipp M. Scholl

A Feasibility Study of Wrist-Worn Accelerometer Based Detection of Smoking Habits

Cigarette smoking is one of the major causes of lung cancer, and has been linked to a large amount of other cancer types and diseases. Smoking cessation, the only mean to avoid these serious risks, is hindered by the ease to ignore these risks in day-to-day life. In this paper we present a feasibility study with smokers wearing an accelerometer device on their wrist over the course of a week to detect their smoking habits based on detecting typical gestures carried out while smoking a cigarette. We provide a basic detection method that identifies when the user is smoking, with the goal of building a system that provides an individualized risk estimation to increase awareness and motivate smoke cessation. Our basic method detects typical smoking gestures with a precision of 51.2% and shows a user-specific recall of over 70% - creating evidence that an unobtrusive wrist-watch-like sensor can detect smoking.

When do you light a fire?: capturing tobacco use with situated, wearable sensors

An important step towards assessing smoking behavior is to detect and log smoking episodes in an unobtrusive way. Detailed information on an individuals consumption can then be used to highlight potential health risks and behavioral statistics to increase the smokers awareness, and might be applied in smoking cessation programs. In this paper, we present an evaluation of two different monitoring prototypes which detect a users smoking behavior, based on augmenting a lighter. Both prototypes capture and record instances when the user smokes, and are sufficiently robust and power efficient to allow deployments of several weeks. A real-world feasibility study with 11 frequently-smoking participants investigates the deployment and adoption of the system, hinting that smokers are generally unaware of their daily smoking patterns, and tend to overestimate their consumption.

Extended explicit semantic analysis for calculating semantic relatedness of web resources

Finding semantically similar documents is a common task in Recommender Systems. Explicit Semantic Analysis (ESA) is an approach to calculate semantic relatedness between terms or documents based on similarities to documents of a reference corpus. Here, usually Wikipedia is applied as reference corpus. We propose enhancements to ESA (called Extended Explicit Semantic Analysis) that make use of further semantic properties of Wikipedia like article link structure and categorization, thus utilizing the additional semantic information that is included in Wikipedia. We show how we apply this approach to recommendation of web resource fragments in a resource-based learning scenario for self-directed, on-task learning with web resources.

Collaborative Semantic Tagging of Web Resources on the Basis of Individual Knowledge Networks

The web is increasingly used as an information source to gain new knowledge but the management of found web pages can be a challenging task. Often social tagging systems are used for resource management. Besides the obvious use of tags --- organizing a collection of web resources --- they support functionalities like sharing resources with other users and recommendation of further possibly relevant web pages. This paper describes a novel application based on an extended tagging concept that can improve resource management and recommendation. Adding semantic information to tags and tagging fragments of web pages instead of whole web pages enhance the possibilities of well-known tagging applications. Individual knowledge networks are the basis of this tagging concept. A first prototype is developed as proof of concept.

Extracting Human Behavior Patterns from Appliance-level Power Consumption Data

In order to provide useful energy saving recommendations, energy management systems need a deep insight in the context of energy consumption. Getting those insights is rather difficult. Either exhaustive user questionnaires or the installation of hundreds of sensors are required in order to acquire this data. Measuring the energy consumption of a household is however required in order to find and realize saving potentials. Thus, we show how to gain insights in the context of energy consumption directly from the energy consumption profile. Our proposed methods are capable of determining the user’s current activity with an accuracy up to 98% as well as the user’s current place in a house with an accuracy up to 97%. Furthermore, our solution is capable of detecting anomalies in the energy consumption behavior. All this is mainly achieved with the energy consumption profile.

Towards Benchmarked Sleep Detection with Wrist-Worn Sensing Units

The monitoring of sleep by quantifying sleeping time and quality is pivotal in many preventive health care scenarios. A substantial amount of wearable sensing products have been introduced to the market for just this reason, detecting whether the user is either sleeping or awake. Assessing these devices for their accuracy in estimating sleep is a daunting task, as their hardware design tends to be different and many are closed-source systems that have not been clinically tested. In this paper, we present a challenging benchmark dataset from an open source wrist-worn data logger that contains relatively high-frequent (100Hz) 3D inertial data from 42 sleep lab patients, along with their data from clinical polysomnography. We analyse this dataset with two traditional approaches for detecting sleep and wake states and propose a new algorithm specifically for 3D acceleration data, which operates on a principle of Estimation of Stationary Sleep-segments (ESS). Results show that all three methods generally over-estimate for sleep, with our method performing slightly better (almost 79% overall median accuracy) than the traditional activity count-based methods.

Cross-lingual recommendations in a resource-based learning scenario

CROKODIL is a platform supporting resource-based learning scenarios for self-directed, on-task learning with web resources. As CROKODIL enables the forming of possibly large learning communities, the stored data is growing in a large scale. Thus, an appropriate recommendation of tags and learning resources becomes increasingly important for supporting learners. We propose semantic relatedness between tags and resources as a basis of recommendation and identify Explicit Semantic Analysis (ESA) using Wikipedia as reference corpus as a viable option. However, data from CROKODIL shows that tags and resources are often composed in different languages. Thus, a monolingual approach to provide recommendations is not applicable in CROKODIL. Thus, we examine strategies for providing mappings between different languages, extending ESA to provide cross-lingual capabilities. Specifically, we present mapping strategies that utilize additional semantic information contained in Wikipedia. Based on CROKODILs application scenario, we present an evaluation design and show results of cross-lingual ESA.

jNode: A sensor network platform that supports distributed inertial kinematic monitoring

Because of the intrinsic advantages of wireless inertial motion tracking, standalone devices that integrate inertial motion units with wireless networking capabilities have gained much interest in recent years. Several platforms, both commercially available and academic, have been proposed to balance the challenges of a small form-factor, power consumption, accuracy and processing speed. Applications include ambulatory monitoring to support healthcare, sport activity analysis, recognizing human group behaviour, navigation support for humans, robots and unmanned vehicles, but also in structural monitoring of large buildings. This paper provides an analysis of the current state-of-the-art platforms in wireless inertial motion tracking and presents a novel open-source and open-hardware hybrid tracking platform that is extensible, low-power, flexible enough to be used for both short- and long-term monitoring and based on a firmware that allows it to be easily adapted after being deployed.

Implementation and Evaluation of a Tool for Setting Goals in Self-regulated Learning with Web Resources

Learning effectively and efficiently with web resources demands distinct competencies in self-organization and self-motivation. According to the theory of Self-Regulated Learning, learning processes can be facilitated and supported by an effective goal-management. Corresponding to these theoretic principles, a goal-management tool has been implemented in an interdisciplinary project. It allows learners to set goals for internet research and assign relevant web resources to them. An evaluation study is presented that focuses on short-term learning episodes and selected results are shown that reinforce the benefits of our approach.

Localization in Wireless networks via Laser scanning and Bayesian compressed sensing

WiFi indoor localization has seen a renaissance with the introduction of RSSI-based approaches. However, manual fingerprinting techniques that split the indoor environment into predefined grids are implicitly bounding the maximum achievable localization accuracy. WoLF, our proposed Wireless localization and Laser-scanner assisted Fingerprinting system, solves this problem by automating the way indoor fingerprint maps are generated. We furthermore show that WiFi localization on the generated high resolution maps can be performed by sparse reconstruction which exploits the peculiarities imposed by the physical characteristics of indoor environments. Particularly, we propose a Bayesian Compressed Sensing (BCS) approach in order to find the position of the mobile user and dynamically determine the sufficient number of APs required for accurate positioning. BCS employs a Bayesian formalism in order to reconstruct a sparse signal using an undetermined system of equations. Experimental results with data collected in a university building validate WoLF in terms of localization accuracy under actual environmental conditions.