An Auto-Encoder Multitask LSTM Model for Boundary Localization

被引:12
作者
Liu, Yu-Ting [1 ]
Chen, Jen-Jee [1 ]
Tseng, Yu-Chee [1 ,2 ,3 ]
Li, Frank Y. [4 ]
机构
[1] Natl Yang Ming Chiao Tung Univ NYCU, Coll Artificial Intelligence, Hsinchu 300093, Taiwan
[2] Acad Sinica, Taipei 11529, Taiwan
[3] Kaohsiung Med Univ, Kaohsiung 80708, Taiwan
[4] Univ Agder UiA, Dept Informat & Commun Technol, N-4898 Grimstad, Norway
关键词
Location awareness; Deep learning; Global Positioning System; Data models; Multitasking; Data collection; Codes; Fingerprint positioning; indoor and outdoor boundary areas; deep learning; auto-encoder and LSTM; multi-sensor and multi-task; real-life experiments;
D O I
10.1109/JSEN.2022.3168253
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
Although positioning in both indoor and outdoor environments has been investigated in-depth individually by various means, little attention has been paid to the intersected areas of both environments and the positioning accuracy in these boundary areas remains concern. For indoor positioning, several recent studies have applied machine learning algorithms based on WiFi signals. However, these methods may not be applicable to solve the boundary localization problem which is faced for positioning in boundary areas around a building. In this paper, we propose a deep learning model which concatenates auto-encoders with LSTM networks to perform multi-sensor multi-task fingerprint positioning. The adopted sensors include WiFi, GPS, cellular (fused with GPS), and magnetometer. In our model, we first employ a dense block-based auto-encoder to extract representative latent codes of fingerprints. Such latent codes are proven to be more distinguishable since they are revertible to their original inputs. Then, a sequence of latent codes are injected into an LSTM network performing three output tasks, responsible for location estimation, most suitable sensor selection, and indoor/outdoor classification, respectively. Extensive real-life experiments are performed on two campuses and the results demonstrate that our model achieves high positioning accuracy with an average Euclidean distance estimation error of about 0.43 meter.
引用
收藏
页码:10940 / 10953
页数:14
相关论文
共 37 条
[1]   WiDeep: WiFi-based Accurate and Robust Indoor Localization System using Deep Learning [J].
Abbas, Moustafa ;
Elhamshary, Moustafa ;
Rizk, Hamada ;
Torki, Marwan ;
Youssef, Moustafa .
2019 IEEE INTERNATIONAL CONFERENCE ON PERVASIVE COMPUTING AND COMMUNICATIONS (PERCOM), 2019,
[2]   UWB Localization in a Smart Factory: Augmentation Methods and Experimental Assessment [J].
Barbieri, Luca ;
Brambilla, Mattia ;
Trabattoni, Andrea ;
Mervic, Stefano ;
Nicoli, Monica .
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, 2021, 70
[3]   AUTO-ASSOCIATION BY MULTILAYER PERCEPTRONS AND SINGULAR VALUE DECOMPOSITION [J].
BOURLARD, H ;
KAMP, Y .
BIOLOGICAL CYBERNETICS, 1988, 59 (4-5) :291-294
[4]  
Chawathe S. S., 2009, P 12 INT IEEE C INT, P1
[5]   WiFi Fingerprinting Indoor Localization Using Local Feature-Based Deep LSTM [J].
Chen, Zhenghua ;
Zou, Han ;
Yang, JianFei ;
Jiang, Hao ;
Xie, Lihua .
IEEE SYSTEMS JOURNAL, 2020, 14 (02) :3001-3010
[6]   Magnetic Field-Based Localization in Factories Using Neural Network With Robotic Sampling [J].
Chiang, Ting-Hui ;
Sun, Zao-Hung ;
Shiu, Huan-Ruei ;
Lin, Kate Ching-Ju ;
Tseng, Yu-Chee .
IEEE SENSORS JOURNAL, 2020, 20 (21) :13110-13118
[7]   An integrated autoencoder-based hybrid CNN-LSTM model for COVID-19 severity prediction from lung ultrasound [J].
Dastider, Ankan Ghosh ;
Sadik, Farhan ;
Fattah, Shaikh Anowarul .
COMPUTERS IN BIOLOGY AND MEDICINE, 2021, 132
[8]  
Dozat T., 2016, Incorporating nesterov momentum into adam
[9]  
Eck Daniel, 2015, IFAC - Papers Online, V48, P34, DOI 10.1016/j.ifacol.2015.08.104
[10]  
Ge XB, 2016, INT CONF SOFTW ENG, P135, DOI 10.1109/ICSESS.2016.7883033