Device-free passive wireless localization system with transfer deep learning method
Recently, device-free passive wireless indoor localization (DFPWL) has attracted great interest due to the widespread deployment of Wi-Fi devices and the rapid growth in demand for location-based services (LBS). The DFPWL fingerprinting approach based on channel state information (CSI) has become the mainstream method since its simple deployment and localization accuracy. It determines the location of the target from the new measurement CSI by collecting a training database that measures the CSI and using a machine learning classifier. However, we have found through experiments that even if the indoor environment does not change, the CSI fingerprint will be different from the CSI fingerprint in the database over time, and most of the CSI-based DFPWL fingerprinting method ignores this. To cope with the reduction in localization accuracy caused by the time-varying characteristic of CSI, we propose a novel transfer deep learning-based DFPWL system in this paper. It uses the CSI extracted from a single link to estimate the location of the target, neither requiring the target to wear any electronic equipment nor deploying a large number of APs and Monitor Devices. Unlike the other traditional CSI-based DFPWL fingerprinting approaches using the pre-processed CSI samples as fingerprints, our system utilizes the transfer deep learning (TDL) method to learn new feature representations from the CSI samples as fingerprints, which can simultaneously minimize the intra-class differences, maximize inter-class differences, and minimize the distribution differences between fingerprint database and test samples. Finally, the KNN algorithm is utilized to compare the test samples and the fingerprint database under the new feature representation to obtain the estimation location of the target. Experiment results show that our system can effectively improve localization accuracy compared to the other state-of-art, and can maintain stable localization accuracy for a long time without reacquiring the CSI fingerprint database.
KeywordsDevice-free indoor localization Channel state information (CSI) Deep learning Transfer learning
This work was supported by the National Natural Science Foundation of China (NSFC) under Grant Nos. 61673310 and 61703324, and the Natural Science Foundation of Shaanxi Province under Grant No. 2019JQ-215.
Compliance with ethical standards
Conflict of interest
We declare that we have no conflict of interest.
- Abadi M, Agarwal A, Barham P et al (2016) Tensorflow: large-scale machine learning on heterogeneous distributed systems. arXiv Prepr arXiv160304467Google Scholar
- Cui Z, Chang H, Shan S, Chen X (2014) Generalized unsupervised manifold alignment. Adv Neural Inf Process Syst 3:2429–2437Google Scholar
- Fernando B, Habrard A, Sebban M, Tuytelaars T (2013) Unsupervised visual domain adaptation using subspace alignment. In: Proceedings of the IEEE International Conference on Computer Vision. pp 2960–2967Google Scholar
- Gong B, Shi Y, Sha F, Grauman K (2012) Geodesic flow kernel for unsupervised domain adaptation. In: Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. pp 2066–2073Google Scholar
- Hu J, Lu J, Tan Y-P (2015) Deep transfer metric learning. In: Proceedings of the IEEE conference on computer vision and pattern recognition. pp 325–333Google Scholar
- Li X, Li S, Zhang D et al (2016) Dynamic-MUSIC: accurate device-free indoor localization. In: UbiComp’16. pp 196–207Google Scholar
- Long M, Wang J, Ding G et al (2013) Transfer feature learning with joint distribution adaptation. In: Proceedings of the IEEE International Conference on computer vision. pp 2200–2207Google Scholar
- Ohara K, Maekawa T, Kishino Y et al (2015) Transferring positioning model for device-free passive indoor localization. In: UbiComp 2015—Proceedings of the 2015 ACM International Joint Conference on Pervasive and Ubiquitous Computing. ACM Press, New York, New York, USA, pp 885–896Google Scholar
- Qian K, Wu C, Yang Z et al (2014) PADS: Passive detection of moving targets with dynamic speed using PHY layer information. In: 2014 20th IEEE International Conference on parallel and distributed systems (ICPADS). IEEE, pp 1–8Google Scholar
- Qian K, Wu C, Yang Z et al (2016) Decimeter level passive tracking with WiFi. In: Proceedings of the Annual International Conference on mobile computing and networking, MOBICOM. pp 44–48Google Scholar
- Sabek I, Youssef M (2013) MonoStream: a minimal-hardware high accuracy device-free WLAN localization systemGoogle Scholar
- Samui P, Roy SS, Balas VE (2017) Handbook of neural computation. Academic Press, CambridgeGoogle Scholar
- Wang X, Wang X, Mao S (2017b) CiFi: Deep convolutional neural networks for indoor localization with 5 GHz Wi-Fi. In: 2017 IEEE International Conference on Communications (ICC). IEEE, pp 1–6Google Scholar
- Wu K, Jiang X, Yi Y et al (2012a) FILA: Fine-grained indoor localization. Proc—IEEE INFOCOM 2210–2218Google Scholar
- Wu K, Xiao J, Yi Y et al (2012c) FILA: fine-grained indoor localization. In: 2012 Proceedings IEEE INFOCOM. pp 2210–2218Google Scholar
- Wu C, Yang Z, Zhou Z et al (2015) PhaseU: real-time LOS identification with WiFi. In: Proceedings—IEEE INFOCOM. pp 2038–2046Google Scholar
- Zhou Z, Yang Z, Wu C et al (2014) LiFi: Line-Of-Sight identification with WiFi. In: Proceedings—IEEE INFOCOM. IEEE, pp 2688–2696Google Scholar