The StoryTeller

Abstract

Due to the recent proliferation of location-based services indoors, the need for an accurate floor estimation technique that is easy to deploy in any typical multi-story building is higher than ever. Current approaches that attempt to solve the floor localization problem include sensor-based systems and 3D fingerprinting. Nevertheless, these systems incur high deployment and maintenance overhead, suffer from sensor drift and calibration issues, and/or are not available to all users. In this paper, we propose StoryTeller, a deep learning-based technique for floor prediction in multi-story buildings. StoryTeller leverages the ubiquitous WiFi signals to generate images that are input to a Convolutional Neural Network (CNN) which is trained to predict loors based on detected patterns in visible WiFi scans. Input images are created such that they capture the current WiFi-scan in an AP-independent manner. In addition, a novel virtual building concept is used to normalize the information in order to make them building-independent. This allows StoryTeller to reuse a trained network for a completely new building, significantly reducing the deployment overhead. We have implemented and evaluated StoryTeller using three different buildings with a side-by-side comparison with the state-of-the-art floor estimation techniques. The results show that StoryTeller can estimate the user's floor at least 98.3% within one floor of the actual ground truth floor. This accuracy is consistent across the different testbeds and for scenarios where the models used were trained in a completely different building than the tested building. This highlights StoryTeller's ability to generalize to new buildings and its promise as a scalable, low-overhead, high-accuracy floor localization system.

Accurate and efficient floor localization with scalable spiking graph neural networks

“…To address this issue, we perform random oversampling to increase the number of samples in minority classes until a balance is reached. Such an operation is also for a fair comparison with SOTA StoryTeller (Elbakly & Youssef, 2020) that adopts the same strategy for data augmentation.…”

“…Since the reference (Elbakly & Youssef, 2020) reported the performances of StoryTeller, TrueStory, and Locus in the Building 0 and 1 of the same UJIIndoorLoc dataset, we directly report their floor identification accuracy mentioned in (Elbakly &Youssef, 2020). For kNN, we implement it by ourselves.…”

“…The fingerprinting approaches, including WiFi fingerprinting and cellular fingerprinting, are very popular due to the wide availability of WiFi and cellular infrastructures. The typical fingerprinting systems for floor localization include SkyLoc (Varshavsky et al, 2007), StoryTeller (Elbakly & Youssef, 2020) and ZeeFI (Gu et al, 2019a). Yet, such methods need a large amount of training data, which increases with the number of floors as well as the area of interest.…”

“…Because deep learning models have the advantages of more powerful representational ability, better generalizability, and unnecessary feature engineering, they have also been applied to floor identification. StoryTeller (Elbakly & Youssef, 2020) uses Convolutional Neural Networks (CNNs) for floor identification, where WiFi signals are first used to generate images that are then fed to a CNN for predicting floors. While StoryTeller is both Access Point (AP)-independent and building-independent, it requires the knowledge of physical building dimensions and 3D locations of APs, which are impractical in some scenarios.…”

“…In this work, FloorLocator is designed to offer several advantages over existing SOTA floor localization methods, including StoryTeller (Elbakly & Youssef, 2020). Specifically, based on graph theory, FloorLocator innovatively utilizes graph topology to learn representations, which enhances the transferability of the model.…”