Analysis of indoor Wi-Fi localization using gaussian process regression and K-nearest neighbor algorithms

Main Article Content

Myo Myint Maw
Hnin Mya Nandar Myo Tint
Sarun Duangsuwan

บทคัดย่อ

Global positioning system (GPS) cannot well localize in indoor environment. Nowadays, indoor Wi-Fi localization is very attractive in positioning and localization research areas because of existing Wi-Fi infrastructure can be used to conduct, so there is no need extra hardware requirements and cost is not expensive. In indoor Wi-Fi localization, the received signal strength indicator (RSSI) fingerprinting played a key role in the access point performance services. This paper proposed the accuracy analysis of indoor Wi-Fi signal based on localization using machine learning approach. Four accept points (APs) were used to measure RSSI and the measured RSSI data were configured as RSSI fingerprint database which was composed of RSSI data of each AP and the position of receiver point. Four corridors which has 1m x 1m has were researched for indoor Wi-Fi localization. 3000 training RSSI and 234 testing RSSI data points were applied in indoor localization. The machine learning algorithms: Gaussian process regression (GPR) and K-nearest neighbor (KNN) approaches were proposed to analyze the accuracy of indoor localization. The results of Wi-Fi localization accuracy were shown using GPR and KNN in indoor environment. The accuracy and mean square error were discussed for indoor Wi-Fi based on GPR and KNN. In this proposed Wi-Fi indoor localization, the accuracy using GPR performed 75%. Moreover, the accuracy using KNN was 83% when K neighbor value was five. When K neighbor value was 10, the accuracy also outperformed 83%. Finally, according to the results, the analysis of corridors using KNN could provide more accuracy than GPR for testing environment in this study.

Article Details

บท
บทความวิจัย

References

Wang C, Kao T, Fang S, et al. Robust Wi-Fi Location Fingerprinting Against Device Diversity Based on Spatial Mean Normalization. Proceedings of Asia-Pacific Signal and Information Processing Association Annual Sumit and Conference (APSIPA ASC ) [Internet]. Kaohsiung, Taiwan; 2013. Available from: http://hb.diva-portal.org/smash/get/diva2:1095554/FULLTEXT01.pdf

Zafari F, Gkelias A, Leung King K, A Survey of Indoor Localization Systems and Technologies, IEEE Communications Surveys and Tutorials, 2019. p 2568 - 99.

Akram BA, Akbar A, Wi-Fi Fingerprinting Based Room Level Indoor Localization Framework Using Ensemble Classifiers. [Internet]. Mehran,2019. DOI: 10.22581/muet 1982.1901.13.

Duvallet F, Tews A D, WiFi Position Estimation in Industrial Environments Using Gaussian Processes. [Internet]. IEEE; 2008: DOI: 10.1109/IROS. 2008.4650910.

Mascharka D, Manley E, Machine Learning for Indoor Localization Using Mobile Phone-Based Sensors[Internet]. 2015. Available from: https://arxiv.org/pdf/1505.06125.pdf

Wu Z, Xu Q, Li J, et al. Passive Indoor Localization Based on CSI and Naive Bayes Classification. [Internet]. IEEE; 2018. DOI: 10.1109/TSMC.2017.2679725.

Christoffer O, Tobias O, The precision of RSSI-fingerprinting based on connected Wi-Fi devices [dissertation]. Högskolan i Borås; 2017.

Amirsoori S, Daud SM, Ahmad N A, et al. WI-FI Based Indoor Positioning Using Fingerprinting Methods (KNN Algorithm) in Real Environment. International[Internet]. Journal of Future Generation Communication and Networking. 2017; 10(9): 23-36 Avialable from: http://dx.doi.org/10.14257/ijfgcn.2017.10.9.03.

Rasmussen CE, Gaussian processes in machine learning, in Advanced Lectures on Machine Learning[Internet]. https://www.cs.ubc.ca/~hutter/EARG.shtml/earg/papers05/rasmussen_gps_in_ml.pdf

Ding X, Gao L, Wang Z, Modified Fingerprinting Algorithm for Indoor Location[Internet], Journal of Communications. 2017; 12(3): 145-51 Available from: http://www.jocm.us/uploadfile/2017/0331/20170331050729488.pdf.