WEDNESDAY, JANUARY 21, 2015
Place: 2D309 (edTech lab), Science Park, Joensuu
Time: 13:15 - 16:00 Finnish time (GMT+2)
13:15 - 14:00, Teemu H. Laine, Assistant Professor, Ajou University, South Korea
Abstract: In this presentation I will demonstrate recent research of the UbiLife research group at Ajou University. Specifically, I will describe how context-awareness and games can be combined to create novel gaming experiences in the domains of education and wellbeing. Three games will be covered: 1) Calory Battle; AR exergame that utilizes location-awareness and augmented reality to promote physical exercise, 2) Running Othello distributed exergame that combines inertial sensors and NFC to enable gameplay between geographically distributed players, and 3) Leometry story-based science learning game with augmented reality. These examples provide the audience with an overview of our research before diving into more specific research topics of context-aware middleware and user context detection.
14:00 - 14:45, Joonas Westlin, Master student, Ajou University, South Korea
Abstract: Developing context-aware applications brings challenges to developers. The developers do not know what kind of devices the users will have, and adding support for many different kinds of devices is time-consuming. Secondly, reasoning about higher level context information (e.g. user activity) is complicated. ManySense is a middleware which facilitates easier development of context-aware applications by providing an abstraction layer which allows querying for context information through query languages. In this presentation I will discuss the aforementioned problems developers face, as well as describe the architecture and features of ManySense. I will also present a demo application which uses ManySense.
14:45 - 15:00 Break
15:00 - 15:45, Jungryul Seo, Doctoral student, Ajou University, South Korea
Abstract: An important component of context-aware computing is how to detect a user's context. In most cases, raw sensor data cannot be used meaningfully to detect the user's context. The essence of context-aware computing is to refine raw sensor data or other low level data in order to determine the user's context based on this refined data. Refining context data is typically based on machine learning and data mining algorithms. In this presentation I will cover basic techniques and algorithms for three aspects of a user's context: step, activity and indoor position. Specifically, I will discuss why these aspects are important in context-aware computing, and how step, activity and indoor positioning data can be inferred based on raw sensor data.
It is possible to follow the presentations online via the Adobe Connect Pro system. If you want to follow the seminar online, please contact Jarkko Suhonen (jarkko.suhonen(at)uef.fi) at least couple of days before the seminar.
For more information about the research seminar, please contact Jarkko Suhonen (jarkko.suhonen(at)uef.fi).