- 【Updated on May 12, 2025】 Integration of CiNii Dissertations and CiNii Books into CiNii Research
- Trial version of CiNii Research Automatic Translation feature is available on CiNii Labs
- Suspension and deletion of data provided by Nikkei BP
- Regarding the recording of “Research Data” and “Evidence Data”
Smartphone Accurate 3-D Localization Using Mirror Image Speakers
Bibliographic Information
- Other Title
-
- 鏡像スピーカを用いたスマートフォン高精度3次元測位手法
Search this article
Description
本稿では,スピーカ2台から送信される音響信号の直接波と反射波を用いたスマートフォンのための新たな高精度3次元測位手法について述べる.我々は,壁と床からの1次反射波を,壁と床に対するスピーカの鏡像体からの信号ととらえる,鏡像スピーカの概念を新たに提案する.また,スマートフォンを水平にもち,2つの搭載マイクロフォンを利用することで実在スピーカおよび鏡像スピーカからの信号を特定する.2.5mの間隔でスピーカを2台設置し,フロアがリフレクティブな廊下とアブソープティブな廊下にて各々12点での測位実験を行った.評価実験を通して,90th-percentileでの測位誤差が37.7mm未満であることと,最大偶然誤差が5.22mm未満であることを確認した.さらに,Dilution of Precision(DoP)やシミュレーションを用いた考察を行い,鏡像スピーカの有用性を明らかにした.
In this paper, we describe a novel high-accuracy 3-D positioning method for a smartphone using direct and reflected acoustic signals from two existing installed speakers. We propose a new concept called Mirror image speaker regarding primary reflected signals from a wall and a floor as signals transmitted from a virtual speaker at a mirror image position of its original one. We hold a smartphone horizontally, and specify a real speaker and mirror image speakers by using two microphones built-in a smartphone. In our experiments, we installed speakers at intervals of 2.5m along a corridor and estimated the smartphone position at several places on a reflective and an absorptive floor. From these results, we obtained 90th-percentile errors of less than 37.7mm and maximum random errors of less than 5.22mm for 3-D positioning. Furthermore, we clarified the usefulness of mirror image speakers using Dilution of Precision (DoP) measures and confirmed it through computer simulations.
Journal
-
- 情報処理学会論文誌
-
情報処理学会論文誌 60 (12), 2314-2324, 2019-12-15
- Tweet
Details 詳細情報について
-
- CRID
- 1050564289020400128
-
- NII Article ID
- 170000181220
-
- NII Book ID
- AN00116647
-
- ISSN
- 18827764
-
- Text Lang
- ja
-
- Article Type
- journal article
-
- Data Source
-
- IRDB
- CiNii Articles
- KAKEN