Document Type
Thesis - Open Access
Award Date
2018
Degree Name
Master of Science (MS)
Department / School
Electrical Engineering and Computer Science
First Advisor
Myounggyu Won
Abstract
A traffic monitoring system (TMS) is an integral part of Intelligent Transportation Systems (ITS) for traffic analysis and planning. However, covering huge miles of rural highways (119,247 miles in U.S.) with a large number of TMSs is a very challenging problem due to the cost issue. This paper aims to address the problem by developing a low-cost and portable TMS called DeepWiTraffic based on COTs WiFi devices. The proposed system enables accurate vehicle detection (counting) and classification by exploiting the unique WiFi Channel State Information (CSI) of passing vehicles. Spatial and temporal correlations of CSI amplitude and phase data are identified and analyzed using a deep learning technique to classify a vehicle into five different types: motorcycle, passenger vehicle, SUV, pickup truck, and large truck (a vehicle with more than three axles according to the FHWA classification). The principal component analysis (PCA) technique is exploited to reduce the dimension of the subcarriers and remove the device specific noise. The CSI phase data of a received signal are preprocessed by applying a linear transformation and the correlations of CSI phase information of multiple subcarriers are taken into account for effective vehicle classification. A convolutional neural network (CNN) is designed to extract optimal features of the preprocessed CSI amplitude and phase data. A huge amount of CSI data of passing vehicles as well as ground truth video data are collected for about 120 hours to validate the performance of the proposed proof-of-concept system. The results show that the average detection accuracy of 99.4%, and the average classification accuracy of 91.1% (Motorcycle: 97.2%, Passenger Car: 91.1%, SUV: 83.8%, Pickup Truck: 83.3%, and Large Truck: 99.7%) can be achieved with a very small cost of less than $1,000.
Library of Congress Subject Headings
Traffic monitoring.
Vehicle detectors.
Wireless communication systems.
Description
Includes bibliographical references
Format
application/pdf
Number of Pages
55
Publisher
South Dakota State University
Recommended Citation
Sahu, Sayan, "Enabling Low Cost WIFI-Based Traffic Monitoring System Using Deep Learning" (2018). Electronic Theses and Dissertations. 2688.
https://openprairie.sdstate.edu/etd/2688
Included in
Digital Communications and Networking Commons, Electrical and Computer Engineering Commons
Rights Statement
