A Novel Positioning Feature for Handling Hardware VarianceProblem.

• Main Authors: Chu-Hsuan Wang, 王楚軒
• Other Authors: Shih-Hau Fang
• Format: Others
• Language: en_US
• Published: 2011
• Online Access: http://ndltd.ncl.edu.tw/handle/55532085398661655932

碩士 === 元智大學 === 電機工程學系 === 99 === The received signal strength (RSS) is the most popular location ngerprinting in Wi-Fi indoor location systems. However, the heterogeneity among the hardware of the devices present challenges to RSS-based location systems. The varying RSSs between training and testing devices would cause a large biased error in location systems. This paper presents a novel positioning feature, named delta-fused principal strength, which uses the RSS and pair-wise delta strength from the access points as inputs to infer the client position in the wireless LAN environment. We extract the useful information into location fi ngerprinting via principal component analysis. Our positioning feature overcomes the hardware-variation problem and provides much higher accuracy compared with other examined techniques. The proposes feature is developed in a real world Wi-Fi environment in localization systems. We implement the di erent positioning features on the same test-bed and compare the results. On-site experimental results demonstrate that DFPS outperforms examined techniques and reduces the 50th and 67th circular error probable by 18.14%-25.52% and 10.79%-31.65%.