A Scalable Framework for Indoor Localization Using Convolutional Neural Networks

Abstract

Indoor localization is a rapidly evolving application domain for enhanced tracking of people and assets in indoor environments such as airports, hospitals, and underground mines. Most prior works in the domain of indoor localization deliver inadequate localization accuracies without expensive infrastructure. Alternatively, methodologies employing inexpensive off-the-shelf devices that are ubiquitous in nature lack consistency in localization quality. Ambient wireless received signal strength indication (RSSI)-based fingerprinting using smart mobile devices is a low-cost approach to the problem. However, creating an accurate “fingerprinting-only” solution remains an open challenge. This chapter presents a novel approach to transform WiFi signatures into images, to create a scalable fingerprinting framework based on convolutional neural networks (CNNs). Our proposed CNN-based indoor localization framework (CNNLOC) is validated across several indoor locales and shows improvements over the best-known prior works, with an average localization error of <2 meters.