Reliable location sensing through multi-sensor fusion, dynamic weighting, and confidence mapping

Naisbitt, Jeffrey D.

Permalink

https://hdl.handle.net/2142/14686 Copy

Description

Title

Reliable location sensing through multi-sensor fusion, dynamic weighting, and confidence mapping

Author(s)

Naisbitt, Jeffrey D.

Issue Date

2010-01-06T16:21:17Z

Director of Research (if dissertation) or Advisor (if thesis)

Campbell, Roy H.

Doctoral Committee Chair(s)

Campbell, Roy H.

Department of Study

Computer Science

Discipline

Computer Science

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

M.S.

Degree Level

Thesis

Keyword(s)

• location-sensing
• tracking
• sensor fusion

Abstract

Ubiquitous computing environments provide multitudes of technologies seamlessly augmented with physical systems to aid users in everyday tasks. In order for these systems to be pervasive and yet imperceptible to the user, they must maintain the location of users, devices, and resources within the room. Real-world systems are dynamic and constantly changing, with mobile people and devices. Current location sensing trends within pervasive systems focus on heterogeneous and hybrid systems that consist of multiple location sensing technologies. Fusing the information from these technologies continues to be a focus of current research. Several problems arise from current approaches to this problem. First, current approaches rely on statically assigned confidences to sensor systems. Obtaining confidence information about sensor technologies in realistic environments remains a difficult challenge. Currently, the analysis of location-sensing technologies occurs under idealized and hypothetical environments, largely ignoring dynamic, environment-specific problems. Current probabilistic approaches improve overall results, but they still lack robustness to the weaknesses inherent in the technologies. Each location sensing system fails to provide reliably accurate results under certain system and environment specific circumstances and scenarios. Due to this, reliable location-awareness requires the fusion of heterogeneous location-sensing systems. We provide a method for dynamically evaluating systems in realistic environments, making it possible to optimize the setup of sensor systems (such as camera placement for vision tracking). Integrating internal sensor error information with multiple location sensor technologies, we provide a system for obtaining reliably accurate location information.

Graduation Semester

2009-12

Permalink

http://hdl.handle.net/2142/14686

Copyright and License Information

Copyright Jeffrey D. Naisbitt

Owning Collections