Signal Fuse Learning Method With Dual Bands WiFi Signal Measurements in Indoor Positioning

Abstract

Today, indoor localization technology based on WiFi signals has become more and more popular and applicable. It not only facilitates people's lives but also creates enormous economic value. However, during the propagation of the WiFi signal, it is easily interfered by obstacles, and the signal fluctuation is significant, resulting in low accuracy of positioning. To overcome these problems, we reduce the influence of environmental factors firstly. Then the positioning accuracy is improved by using the SVM model to distinguish the NLOS or LOS environment and employing the capsule networks to derive the users' positions with the WiFi 2.4G and 5G signals. As we all know, the WiFi 2.4G signal has excellent penetrability and is less affected by obstacles, while the WiFi 5G signal has excellent stability and small fluctuations. Therefore, we use the advantages of these two kinds of signals to derive the optimal suggestion by the capsule neural network, which is the learning system with minimum data sets needed. The experimental results show that the positioning effect of the two signals simultaneously is better than the positioning effect of a single signal. We also compare with the traditional indoor positioning methods and use the simulation data to carry out the robustness test, and the positioning accuracy reached 0.99 m in the field environment finally.INDEX TERMS Indoor localization, NLOS and LOS channel propagation condition, WiFi 2.4G and WiFi 5G, SVM, capsule network.

A Survey of Machine Learning in Pedestrian Localization Systems: Applications, Open Issues and Challenges

“…In DL architectures for pedestrian localization, classification techniques could be used to improve the accuracy through a previous site detection or feature classification in the fingerprint. The authors in [56], [61], [67], [69] and [83] use CNN to improve the pedestrian localization system performance. The authors in [69] use continuous wavelet transforms as features to construct the fingerprint.…”

“…Then, they use CNN to estimate the room and the location of a pedestrian. In [67], the authors present a novel system in which they use WiFi dual band fingerprint. They use SVM to line of sight (LOS) and non-LOS (NLOS) detection, and to select the work band.…”

Hybrid Approach for Indoor Localization Using Received Signal Strength of Dual-Band Wi-Fi

“…Some studies used dual-band RSS for fingerprint-based positioning [ 32 , 33 , 34 , 35 , 36 , 37 ]. Consequently, the accuracy of these methods depends on site surveys, and vulnerable to environmental changes.…”

WiFi Signal Strength Degradation Over Different Building Materials

“…Apart from speed, WiFi signal strength also affects the WiFi penetration rate. For example, 2.4GHz and 5GHz WiFi, where the penetration rate of 2.4GHz WiFi signal is higher than 5GHz (Own, C. M., et al, 2019).…”