By Topic

Discriminant Minimization Search for Large-Scale RF-Based Localization Systems

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$31 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

2 Author(s)
Sheng-Po Kuo ; Telcordia Appl. Res. Center, Taipei, Taiwan ; Yu-Chee Tseng

In large-scale fingerprinting localization systems, fine-grained location estimation and quick location determination are conflicting concerns. To achieve finer grained localization, we have to collect signal patterns at a larger number of training locations. However, this will incur higher computation cost during the pattern-matching process. In this paper, we propose a novel discriminant minimization search (DMS)-based localization methodology. Continuous and differentiable discriminant functions are designed to extract the spatial correlation of signal patterns at training locations. The advantages of the DMS-based methodology are threefold. First, with through slope of discriminant functions, the exhaustive pattern-matching process can be replaced by an optimization search process, which could be done by a few quick jumps. Second, the continuity of the discriminant functions helps predict signal patterns at untrained locations so as to achieve finer grained localization. Third, the large amount of training data can be compressed into some functions that can be represented by a few parameters. Therefore, the storage space required for localization can be significantly reduced. To realize this methodology, two algorithms, namely, Newton-PL and Newton-INT, are designed based on the concept of gradient descent search. Simulation and experiment studies show that our algorithms do provide finer grained localization and incur less computation cost.

Published in:

Mobile Computing, IEEE Transactions on  (Volume:10 ,  Issue: 2 )