CNN Localization using AP Inverse Position Estimation

Abstract

This contribution focuses on indoor localization by Finger Print method using RSSI of wireless LAN access point (AP). In recent years, there are lively animated discussions on the localization methods using Deep Learning. We proposed Finger Print based on the Convolutional Neural Network (CNN). Establish the adjacency relationship among APs as a two-dimensional model and use it to make the CNN model for Finger Print localization. In order to confirm the improvement of the localization accuracy by this proposal, we verified by experimental data.