DQ-WiAPDoM: A DQN-based AP Deployment Optimization Method for Wi-Fi FTM Positioning

Zihao Liu,Han Li,Bo Gao,Ke Xiong,Pingyi Fan
DOI: https://doi.org/10.1109/ICSTSN57873.2023.10151521
2023-01-01
Abstract:This paper focuses on how to optimize the deployment location of the anchor Wi-Fi access points (APs) in order to enhance the indoor positioning performance of the finetime measurement (FTM) protocol. Compared to conventional Wi-Fi positioning, FTM protocol is able to achieve relatively good indoor positioning performance since it estimates distance between a sender and its receiver with the flight duration of Wi-Fi signals. However, the FTM-based positioning accuracy in NLoS environments is still not high, which restricts its application in indoor scenarios. To this end, we formulate an optimization problem to maximize the total number of LoS paths by optimizing the locations of the anchor APs. To solve the problem, we propose a Deep Q-Network (DQN)-based AP deployment optimization method (DQ-WiAPDoM), which utilizes DQN to optimize the locations of APs such that the total number of LoS paths over all covered area is maximized. Experimental results show that DQ-WiAPDoM converges well and is able to reduce the average positioning error efficiently.
What problem does this paper attempt to address?