Sian Lun Lau

Supporting patient monitoring using activity recognition with a smartphone

In our work we envision the use of context awareness to complement and improve telemedicine services. One concrete example is the recognition of physical activity (PA) for therapies and treatments. The usage of context can provide newer services by assisting medical professionals and patients to obtain implicit information. We introduce the Context Aware Remote Monitoring Assistant (CARMA), an application implemented on the MATRIX telemedicine middleware platform. It enables activity recognition for patients using non-obtrusive devices. In this paper, a movement recognition approach is investigated using a smartphone with a built-in accelerometer. Experiments were carried out to compare the influences of classification algorithms, features and combinations for feature extraction conditions on the recognition accuracy. The obtained results indicated that the approach is viable and shows much potential.

An approach to Collaborative Context Prediction

Context prediction approaches forecast future contexts based on known context patterns to adapt e.g., services in advance. In the case of the users context history not providing suitable context information for the observed context pattern, to the best of our knowledge context prediction algorithms will fail to forecast the appropriate future context. To overcome the gap of missing context information in the users context history, we propose the Collaborative Context Prediction (CCP) approach. CCP utilises the collaborative characteristics of existing recommendation systems of social networks. To evaluate the CCP method an experimental comparison of the proposed method against the local Alignment context predictor is carried out.

DCCLA: Automatic Indoor Localization Using Unsupervised Wi-Fi Fingerprinting

People spend most of their time in a few significant places and often indoors in a small number of select rooms and locations. Indoor localization in terms of a users current place, related to a users daily life, routines or activities, is an important context. We implemented an automatic approach DCCLA Density-based Clustering Combined Localization Algorithm to automatically learn the Wi-Fi fingerprints of the significant places based on density-based clustering. In order to accommodate the influence of the signal variation, clustering procedure separately works on a list of RSSIs Received Signal Strength Indicators from each AP Access Point. In this paper, the approach is experimentally investigated in a laboratory setup and a real-world scenario in an office area with adjacent rooms, which is a key challenge to distinguish for place learning and recognition approaches. From these experiments, we compare and identify the most suitable parameters for the unsupervised learning.

Pedestrian Movement Recognition for Radio Based Collision Avoidance: A Performance Analysis

In todays increasingly mobile society, car manufacturers and suppliers worldwide face the imperative requirement and have already been working to reduce the number of people killed in traffic accidents. In this context pedestrian protection respectively collision avoidance between cars and pedestrians is of vital importance. Several approaches exist to detect pedestrians but only a few of them are able to work without line of sight and/ or only have a very brief view of the movement history of the pedestrian. However, radio communication based systems which exchange GPS based position information of the pedestrian and the car might have a detailed view to the movement history but the update frequency of GPS positions provided by the mobile phone reduces the speed of the system and might induce false alarms. In this paper, the additional use of pedestrian movement recognition using an acceleration sensor implemented in a mobile phone in the area of collision avoidance between cars and pedestrian is presented and a performance analysis is carried out using the results of an experiment. Finally, the benefits of this approach to the reduction of false alarms and detection of endangered pedestrians are discussed.

DAGR DAG Based Context Reasoning: An Architecture for Context Aware Applications

Pervasive computing includes the vision of small but intelligent devices around us. While the hardware of such devices is available in the market for years,they are still not intelligent in our every day life. For the success of the idea of pervasive computing a unified programming model is needed. Without a standard way of composing intelligent applications for pervasive computing, especially context aware applications to minimize user input, and without standard communication protocols it will take a longtime till before we will see those applications in the market. To establish such standards we first need to be clear of the requirements and how to deal with those requirements in the best way. In this paper we present three important requirements. Moreover we present anew framework DAGR (Directed acyclic graph based context reasoning) to address these requirements.

3 rd workshop on recent advances in behavior prediction and pro-active pervasive computing

The 2nd Workshop on recent advances in behavior prediction and pro-active pervasive computing focuses on contributions that target recent challenges of context prediction and on applications of context prediction. The main challenges are a lack of benchmarks and common data sets, as well as a lack of development frameworks and that the main focus of context prediction still remains location prediction. Since context prediction is a key requirement to enable proactive applications, the workshop aims to intensify the discussion about the state and direction of context prediction research and to facilitate collaboration among research groups focusing on context prediction.

Evaluation of the Influence of Time Synchronisation on Classification Learning Based Movement Detection with Accelerometers

Context awareness is an important enabler of future applications in ubiquitous environments and has the potential to free the user from the control loop in a situation where he faces ever more connected devices. The context of a user can be accessed using data gathered from the users personal devices and devices in his proximity. In this paper we examine the relation between time synchronisation of devices and context reasoning accuracy. For an application where we detect different activities of an office worker from acceleration data measured at the wrists, we show that insufficient time synchronisation leads to a significant decrease of reasoning accuracy. We show this for six commonly used classification learning algorithms that otherwise offer high accuracy provided devices are time synchronised. We also analyse the reasons that causes a lack of time synchronisation and why known synchronisation protocols are not applicable in ubiquitous environments. To solve the problems imposed by insufficient time synchronisation we propose a conceptual change in the communication paradigm used in architectures for context awareness.

Implementation of a User-Centric Context-Aware Playground

Since the introduction of the ubiquitous computing vision by Mark Weiser, we observed quite some research from various R&D projects as well as groups that focused on different aspects in the area of context-awareness. However, there is yet no result which we can easily apply in a real environment. The context-aware service development and deployment process seldom focused on reusability and realistic implementation issues. We believe these issues can be tackled by proposing a complete context-aware system package. In this package both developers and users will be supported. This idea is realised at our department as a context-aware playground. This paper explains the observation on current work and the shortcomings, and proposes how the context-aware playground approach can be a solution.

A Toolchain for Context Recognition: Automating the Investigation of a Multitude of Parameter Sets

A persons context data can be used for a multitude of applications, such as energy management or health care. Common context recognition approaches rely on several factors, such as the sensor set, features, or the context modeling algorithm. Discovering the recognition performances of different parameter setting combinations is a complex, time-consuming, and error-prone task. To support the context recognition research, we present the Context Recognition Assistance Tool (CRAT). The Context Recognition Assistance Tool assist by automatically conducting the evaluation for a multitude combination of parameter settings and clearly presents the findings. Using the CRAT, we investigate to what degree five parameters influence the recognition accuracy. To support the research in the field of context recognition, the CRAT is publicly available.

Web-based Interactive Form Generator for public Kiosks

Self-service Kiosks are found in many public areas such as banks, retail offices, hospitals etc. The creation and deployment of forms for a kiosk usually requires programming and this is a task performed by programmers with technical know-how. This paper presents Web-based Interactive Form Generator (WebIFGen), a prototype system. WebIFGen is a web-based system that allows creation and deployment of forms without any technical know-how. This is achieved by abstracting essential components and processed identified from common kiosk systems. The paper elaborates the design and implementation of the system and discusses on the lesson learned.