An Interacting Multiple-Model-Sets Estimation Method for Effective TOA Wireless Positioning

• Main Authors: I-Ching Huang, 黃亦靖
• Other Authors: Tan-Jan Ho
• Format: Others
• Language: zh-TW
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/87944051465369394705

碩士 === 中原大學 === 電機工程研究所 === 101 === This thesis presents an interacting multiple-model-sets estimation method for wireless location estimation. In other studies, they discuss with no non-line of sight (NLOS) bias or with fixed NLOS bias. In this thesis, the NLOS bias is time-varying, which can coincide with actual situations in practice. Based on a partition of the NLOS bias range, we can construct model sets and switching conditions between model sets. On this basis, we develop our algorithm. The proposed algorithm uses interacting multiple-model sets and adjusts innovations to remove the NLOS bias. Thanks to this switching between model sets, LOS and NLOS cases can be identified more accurately. This thesis uses time of arrival (TOA) data as measurements. Simulations show that the proposed method performs better than other methods in city environments with varying NLOS bias.