Publication Type

Conference Proceeding Article

Version

Postprint

Publication Date

3-2011

Abstract

Proliferation of mobile applications in unpredictable and changing environments requires applications to sense and act on changing operational contexts. In such environments, understanding the context of an entity is essential for adaptability of the application behavior to changing situations. In our view, context is a high-level representation of a user or entity’s state and can capture activities, relationships, capabilities, etc. Inherently, however, these high-level context measures are difficult to sense directly and instead must be inferred through the combination of many data sources. In pervasive computing environments where this context is of significant importance, a multitude of sensors is already being embedded in the environment to provide streams of low-level sensor data about the environment and the entities present in that environment. A key challenge in supporting context-aware applications in these environments, therefore, is supporting energy-efficient determination of multiple (potentially competing) high-level context measures simultaneously using data from low-level sensor streams. In this paper, we first highlight the key challenges that distinguish the multi-context determination problem from single context determination and then develop our framework and practical implementation to account for them. Our model captures the tradeoff between the accuracy of estimating multiple context measures and the overhead incurred in acquiring the necessary sensor data. Given a set of required contexts to determine, we develop a multi-context search heuristic to compute both the best set of sensors to contribute to context determination and parameters of the sensing tasks. Our algorithm’s goal is to satisfy the applications’ specified needs for accuracy at a minimum cost. We compare the performance of our heuristic approach with a brute-force approach for multi-context determination.Experimental results with SunSPOT sensors demonstrate the potential impact of this approach.

Keywords

context-awareness, multi-context recognition, streaming multi-modal sensors

Discipline

Software Engineering

Research Areas

Software Systems

Publication

Proceedings of IEEE International Conference on Pervasive Computing and Communications

First Page

63

Last Page

73

ISBN

9781424495283

Identifier

10.1109/PERCOM.2011.5767596

Publisher

IEEE

City or Country

Seattle, WA

Creative Commons License

Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.

Additional URL

http://dx.doi.org/10.1109/PERCOM.2011.5767596

Comments

One of 3 Papers Nominated for Mark Weiser Best Paper Award

Share

COinS