LoT: A Transformer-Based Approach Based on Channel State Information for Indoor Localization

被引:3
作者
Li, Wei [1 ]
Meng, Xiangxu [2 ]
Zhao, Zheng [2 ]
Liu, Zhihan [2 ]
Chen, Chuhao [2 ]
Wang, Huiqiang [2 ]
机构
[1] Harbin Engn Univ, Coll Comp Sci & Technol & Modeling & Emulat, E Government Natl Engn Lab, Harbin 150001, Heilongjiang, Peoples R China
[2] Harbin Engn Univ, Coll Comp Sci & Technol, Harbin 150001, Heilongjiang, Peoples R China
关键词
Channel state information (CSI); fingerprint-based localization; indoor positioning; location-based services (LBSs); transformer;
D O I
10.1109/JSEN.2023.3318835
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
Fingerprint-based localization has gained significant commercial value in recent years and have brought new opportunities to location-based services in indoor environments. Existing fingerprint indoor positioning studies have focused on various neural networks (e.g., convolutional neural networks (CNNs), recurrent neural networks (RNNs), and hybrid approaches). However, these methods are limited by their inherent defects, such as local trapping in CNNs and gradient extinction in RNNs, which result in low localization accuracy and high training overhead. In this article, we devise a deep learning localization method, called localization transformer (LoT), to further enhance the indoor localization performance. Then, to address the problem of information loss caused by the input of the channel state information (CSI) matrix into LoT, we propose a trend-cyclical decomposition padding method that performs a trend-cyclical decomposition of the raw data to pad the dimensions. In addition, a rectangular patching strategy is designed to "patch" any shape of CSI matrix, so that the CSI matrix can be fed into a self-attention based learning module. Finally, by using a fully connected network (FCNN) to map the learned features, the 3-D coordinates can be estimated accurately. Experiments were conducted in indoor and urban canyon scenarios with three signal-to-noise ratios (SNRs). The experimental results indicate the mean absolute errors of LoT in indoor and urban scenes is 0.18 and 0.48 m, respectively, which shows that it outperforms the CSI fingerprint-based methods regarding localization accuracy.
引用
收藏
页码:28205 / 28219
页数:15
相关论文
共 51 条
[1]  
Abdallah A., 2022, PROC EUR C ANTENNAS, P1
[2]  
ABDELNAIEM AM, 2021, POLYM BULL, V78, P3995, DOI DOI 10.1007/S00289-020-03301-5
[3]   3D building interior modelling for navigation in emergency response applications [J].
Boguslawski, Pawel ;
Zlatanova, Sisi ;
Gotlib, Dariusz ;
Wyszomirski, Michal ;
Gnat, Milosz ;
Grzempowski, Piotr .
INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION, 2022, 114
[4]   TOA positioning for uplink cooperative NOMA in 5G networks [J].
Celik, G. ;
Celebi, H. .
PHYSICAL COMMUNICATION, 2019, 36
[5]   RWKNN: A Modified WKNN Algorithm Specific for the Indoor Localization Problem [J].
Chen, Guokai ;
Guo, Xiye ;
Liu, Kai ;
Li, Xiaoyu ;
Yang, Jun .
IEEE SENSORS JOURNAL, 2022, 22 (07) :7258-7266
[6]   Soil salinity inversion based on differentiated fusion of satellite image and ground spectra [J].
Chen, Hongyan ;
Ma, Ying ;
Zhu, Axing ;
Wang, Zhuoran ;
Zhao, Gengxing ;
Wei, Yanan .
INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION, 2021, 101
[7]   Graph-Based Radio Fingerprint Augmentation for Deep-Learning-Based Indoor Localization [J].
Chen, Yuxuan ;
Li, Guobing ;
Tan, Yifeng ;
Zhang, Guomei .
IEEE SENSORS JOURNAL, 2023, 23 (06) :6074-6084
[8]   MaMIMO CSI-based positioning using CNNs: Peeking inside the black box [J].
De Bast, Sibren ;
Pollin, Sofie .
2020 IEEE INTERNATIONAL CONFERENCE ON COMMUNICATIONS WORKSHOPS (ICC WORKSHOPS), 2020,
[9]   A TDOA and PDR Fusion Method for 5G Indoor Localization Based on Virtual Base Stations in Unknown Areas [J].
Deng, Zhongliang ;
Zheng, Xinyu ;
Zhang, Chongyu ;
Wang, Hanhua ;
Yin, Lu ;
Liu, Wen .
IEEE ACCESS, 2020, 8 :225123-225133
[10]  
Dev CSGN, 2020, 2020 IEEE 3RD 5G WORLD FORUM (5GWF), P25, DOI 10.1109/5GWF49715.2020.9221379