In-Air Handwriting by Passive Gesture Tracking Using Commodity WiFi

被引：23

作者：

Han, Zijun ^{[1
,2
,3
]}

Lu, Zhaoming ^{[1
,2
,3
]}

Wen, Xiangming ^{[1
,2
,3
]}

Zhao, Jingbo ^{[1
,2
,3
]}

Guo, Lingchao ^{[1
,2
,3
]}

Liu, Yue ^{[1
,2
,3
]}

机构：

[1] Beijing Univ Posts & Telecommun, Sch Informat & Commun Engn, Beijing 100876, Peoples R China

[2] Beijing Univ Posts & Telecommun, Beijing Key Lab Network Syst Architecture & Conve, Beijing 100876, Peoples R China

[3] Beijing Univ Posts & Telecommun, Beijing Lab Adv Informat Networks, Beijing 100876, Peoples R China

来源：

IEEE COMMUNICATIONS LETTERS | 2020年 / 24卷 / 11期

基金：

中国国家自然科学基金;

关键词：

Tracking; Wireless fidelity; Antennas; Transceivers; Calibration; Dynamics; Wireless communication; Channel state information; human-computer interaction; handwriting; gesture tracking;

D O I：

10.1109/LCOMM.2020.3007982

中图分类号：

TN [电子技术、通信技术];

学科分类号：

0809 ;

摘要：

Recent years have witnessed the great potential of adopting Channel State Information (CSI) for human-computer interaction by gestures. However, most current solutions either depend on specialized hardware or demand priori learning of wireless signal patterns, which face critical downsides in availability, reliability and extensibility. Hence this letter presents AirDraw, a novel learning-free in-air handwriting system by passive gesture tracking using only three commodity WiFi devices. First, we denoise CSI measurements by the ratio between two close-by antennas, and further separate the reflected signal from noise by performing Principal Component Analysis. Besides, we propose a robust signal calibration algorithm for tracking correction by eliminating the static components unrelated to hand motion. The prototype of AirDraw is fully realized and evaluated in real scenario. Extensive experiments yield that AirDraw can track user's hand trace with a median error lower than 2.2 cm.

引用

页码：2652 / 2656

页数：5

共 15 条

[1] See Through Walls with Wi-Fi! [J].