Wi-Fi based indoor location positioning employing random forest classifier

Location positioning in indoor environments is a major challenge. Various algorithms have been developed over years to address the problem of indoor positioning. One of the most cost effective choice for indoor positioning is based on received signal strength indicator (RSSI) using existing Wi-Fi networks in commercial and/or public areas. This solution is infrastructure-free and offers meter-range accuracy. In this paper, machine learning approaches including k-nearest neighbor (k-NN), a rules-based classifier (JRip), and random forest have been investigated to estimate the indoor location of a user or an object using RSSI based fingerprinting method. Experimental measurements were carried out using 1500 reference points with received RSSIs of 86 installed APs in the second floor of Centre for Engineering Innovation (CEI) building at the University of Windsor. The results indicate that the random forest classifier presents the best performance as compared to k-NN and JRip classifiers with positioning accuracy higher than 91%.

[1]  Wenyu Liu,et al.  Indoor Localization Based on Curve Fitting and Location Search Using Received Signal Strength , 2015, IEEE Transactions on Industrial Electronics.

[2]  Tacha Serif,et al.  Indoor location detection with a RSS-based short term memory technique (KNN-STM) , 2012, 2012 IEEE International Conference on Pervasive Computing and Communications Workshops.

[3]  Aaron Harwood,et al.  Optimizing indoor location recognition through wireless fingerprinting at the Ian Potter Museum of Art , 2014, 2014 International Conference on Indoor Positioning and Indoor Navigation (IPIN).

[4]  Ahmad Hatami.  Application of Channel Modeling for Indoor Localization Using TOA and RSS , 2006 .

[5]  Yong Liu,et al.  Fingerprint-based location positoning using improved KNN , 2012, 2012 3rd IEEE International Conference on Network Infrastructure and Digital Content.

[6]  Rosdiadee Nordin,et al.  Recent Advances in Wireless Indoor Localization Techniques and System , 2013, J. Comput. Networks Commun..

[7]  Per Enge,et al.  Special Issue on Global Positioning System , 1999, Proc. IEEE.

[8]  Mina Maleki,et al.  A RSS-based fingerprinting method for positioning based on historical data , 2010, Proceedings of the 2010 International Symposium on Performance Evaluation of Computer and Telecommunication Systems (SPECTS '10).

[9]  Xu Zhan,et al.  Indoor location algorithm research based on neural network , 2014, 2014 12th International Conference on Signal Processing (ICSP).

[10]  Noah Pritt,et al.  Indoor positioning with maximum likelihood classification of Wi-Fi signals , 2013, 2013 IEEE SENSORS.

[11]  Limin Xiao,et al.  Measurement-based RSS-multipath neural network indoor positioning technique , 2014, 2014 IEEE 27th Canadian Conference on Electrical and Computer Engineering (CCECE).

[12]  Samih Eisa,et al.  Removing useless APs and fingerprints from WiFi indoor positioning radio maps , 2013, International Conference on Indoor Positioning and Indoor Navigation.

[13]  R. Mautz.  Indoor Positioning Technologies , 2012 .

[14]  Kaveh Pahlavan,et al.  Super-resolution TOA estimation with diversity for indoor geolocation , 2004, IEEE Transactions on Wireless Communications.