CommuterScanner: Towards Smart Indoor Positioning Systems in Urban Transportation

Published In

2019 11th International Conference on Communication Systems & Networks (COMSNETS)

Document Type

Citation

Publication Date

5-2019

Abstract

With the rapid growth of modern cities, public transportation systems require smart planning to provide effective and competitive services to the daily commuters. Due to the presence of Wireless LAN Network (WLAN), Wi-Fi access is available to commuters during their daily commute. The overarching goal of this work is to leverage infrastructure-based indoor positioning systems (IIPS) deployed at train stations to enable several smart transportation use cases by analyzing commuter traffic. Specifically, in this work we address identification of whether a user is in-train or on-platform by utilizing two types of passively sensed Wi-Fi data, namely, received signal strength (RSSI) and phase vectors (AoA) measured at the deployed access points from data received from mobile devices. We conduct structured analysis of each data source, to identify features that distinguish on-platform and in-train devices. OurCommuterScanner solution achieves up to 90% accuracy using random forest model. Our solution works for a variety of deployments including APs with RSSI-only or RSSI + AoA capabilities and irrespective of if the device is connected to the Wi-Fi.

Description

© Copyright 2019 IEEE - All rights reserved.

DOI

10.1109/COMSNETS.2019.8711465

Persistent Identifier

https://archives.pdx.edu/ds/psu/29346

Share

COinS