VLC Positioning by DNN via WkNN in Indoor Environment

Abstract

With the recent development of mobile communication systems, the rapid development of smart devices and internet of things (IoT) technology, and the expansion of the field of application is taking place. One of these application fields is location-based services (LBS). LBS is a technology that provides various services based on a user’s location. However, it is very important to accurately position the user to provide such a service. In the case of an outdoor environment, relatively high positioning accuracy may be provided through a global positioning system (GPS) or the like. However, the application of the GPS is limited in an indoor environment due to problems such as propagation loss. Therefore, in this paper, we propose a method for accurately positioning a user’s location in an indoor environment based on visible light communication (VLC) and artificial intelligence (AI) technology. The proposed scheme proceeds as follows. First, a fingerprinting database is constructed based on the channel characteristics of VLC. Next, the approximate location of the user is obtained by applying a weighted k-nearest neighbor (WkNN). Thereafter, the received signal strength (RSS) value between each access point (AP) and the user equipment (UE) and the user’s approximate location derived through WkNN are used as inputs to the deep neural network (DNN) model to perform learning. The trained DNN model outputs the actual user’s location. Through the simulation results, it can be confirmed that the proposed scheme in this paper achieves precise positioning accuracy compared to the existing scheme.