Home

Home » Research » Future Internet » Context Recognition by User Situation Data Analysis (CONTEXT)


Context Recognition by User Situation Data Analysis (CONTEXT)

The Context project studies characterization and analysis of information about user context and its use in proactive adaptivity. In mobile and ubiquitous applications and systems, reacting to user context is a key component of proactivity: changes in the user’s situation are rapid and they are strongly reflected in the user’s needs and preferences.

The project focuses on the utilization of user context: how does the context reflect the user's motivations, how to make automatic inferences about the contexts, and how to characterize contexts to users and design user interaction about contexts? These questions are considered in the framework of an example application: mobile communication, a representative ubiquitous application whose usability greatly depends on how context-sensitively communication decisions are managed. The project has adopted a multidisciplinary approach where the research problems are approached by qualitative user studies, data analysis algorithm development, and empirical testing in a prototype environment. The key results are (1) methods for utilizing qualitative user situation descriptions in the development of context-sensitive applications, (2) algorithms for context analysis and characterization, and (3) models for user interaction about context. The project has produced a prototype of a context-sensitive mobile communication application.

The project started in November 2002 with qualitative user studies aimed at revealing how users interpret context descriptions and generally context structures. A humanistic research strategy has been deployed to develop a working prototype of proactive context-communication, called ContextPhone, running on a Nokia Series 60 mobile phone. At the same time, ethnomethodological studies by the research group revealed how social and temporal contexts affect interaction and how contexts could be modelled in proactive computers. The prototype has been subjected to a series of longitudinal field studies in 2004 to examine its effects on group communication behavior.

Released in open source, the software has been widely used also outside HIIT and Finland. See http://www.cs.helsinki.fi/group/context/#sw for more information.

In several field studies of ContextPhone carried out during 2005, mobile awareness information was found to have its main supportive functions in coordination of group action, self-expression, and one-to-one companionship. Field experiments carried out together with Nokia Research Center operationalized and measured fragmentation of attention in human-computer interaction in mobile situations. Continuing from this, a doctoral dissertation (of Antti Oulasvirta, University of Helsinki, Department of Psychology) examined and described how human memory systems coordinate the flow of information during interrupted task processing.

Several publications summarized a humanistic, constructive approach to context that problematized the naive realistic approach prominent among engineering approaches to context-awareness. The key idea of the constructive approach is to provide sensor information for the users instead of using it as a basis for automated proactive actions.

The project ended on 31.12.2005. The work will continue in a follow-up project Context Cues, accepted by Academy of Finland for 1.1.2006-31.12.2009 in November 2005.

People

Research groups

Duration

1.11.2002-31.12.2005

Volume

15 person years

Funding

The Academy of Finland

Synopsis

This project studies characterization and analysis of information about user's context and its use in proactive adaptivity. In mobile and ubiquitous applications and systems, reacting to user context is a key component of proactivity: changes in the user's situation are rapid and they are strongly reflected to the user's needs and preferences.

We focus on the utilization of user context: how does the context reflect user's motivations, how to make automatic inferences about the contexts, and how to characterize contexts to users and design user interaction about contexts? These questions are considered in the framework of an example application: mobile messaging, a representative ubiquitous application whose usability greatly depends on how context-sensitively communication is managed. We adopt a multidisciplinary approach: qualitative user studies, data analysis algorithm development, and empirical testing in a prototype environment are methods we will use.

Publications

  • Kankainen, A., & Tiitta, S. (2003). Exploring everyday needs of teenagers related to context-aware mobile services. Proceedings of HFT 2003, Berlin, Germany, 19-26.
  • Tamminen, S., Oulasvirta, A., Toiskallio, K., & Kankainen, A. (2003). Understanding mobile contexts. Proceedings of Mobile HCI 2003, Udine, Italy, 18-35. A revised version submitted to Personal and Ubiquitous Computing.
  • Oulasvirta, A. (2003). Developing future technologies that support social practices and community building. Psykocenter Symposium, Jyväskylä, Nov 3.-4., 2003, 177-186.
  • Kurvinen, E., & Oulasvirta, A. (in press). Towards socially aware pervasive computing: a turntaking approach. Accepted to 2nd International IEEE Conference on Pervasive Computing and Communications, Orlando, Florida, March 2004.
  • Oulasvirta, A. (accepted). Discovering potentials for technology beyond the desktop: The humanistic research strategy. ACM CHI 2004, full paper.
  • Oulasvirta, A., & Tamminen, S. (submitted). Temporal tensions in human-computer interaction.
  • Rantanen, M., & Oulasvirta, A. (in preparation). Simulation approach to ubiquitous computing.

Literature reviews

See also www.cs.helsinki.fi/group/context/ for further information, publications, and links to open source software published by the project.