Honors Program Theses

Award/Availability

Open Access Honors Program Thesis

First Advisor

Dheryta Jaisinghani, Honors Thesis Advisor

Keywords

Machine learning; Electromagnetic interference--Detection; Internet of things;

Abstract

IoT (Internet of Things) devices have become increasingly popular in recent years while WiFi continues to serve as primary network provider indoors. With the advancements in technology, the networks of IoT devices continue to weave closely with indoor WiFi network deployments. Both kinds of these networks primarily operate in 2.4 GHz ISM Band (though latest WiFi standards can operate in 5 GHz and 60 GHz bands, too). With the multitude of tiny IoT sensors being deployed indoors alongside operational WiFi networks, severe interference scenarios cannot be ruled out. As a result of this interference, performance of WiFi networks is bound to suffer considerably.

While the applications and use cases of IoT networks, such as smart buildings, are well appreciated, their co-existence with WiFi networks should be judiciously planned. Although there exists several latest works that study the impact of WiFi networks on the performance of IoT networks, there is a dearth of works that analyze the impact of different IoT networks on the WiFi networks. In this work, we consider major IoT protocols, namely Zigbee, OpenThread, Bluetooth, and Bluetooth Low Energy, and study their impact on the co-deployed WiFi network. We conduct benchmarking experiments for TCP and UDP traffic for different permutations and combinations of these protocols. Our major findings indicate that major IoT protocols can lead to decreased WiFi bandwidth for both TCP and UDP traffic, with our experiments showing reductions of up to 27% in TCP bandwidth and 45% in UDP bandwidth.

One of the aims of this work is to assist network administrators in helping to diagnose the cause of reduced performance of a WiFi network. Off-the-shelf WiFi controllers, such as those from Cisco and Aruba, provide some limited capabilities for diagnosing issues of this kind. With this work, we train and test machine learning models that can detect the patterns in WiFi MAC Layer traffic signifying the performance drop due to IoT interference. Once a cause is established, network administrators can then take appropriate actions to improve the performance of the WiFi network. We test machine learning algorithms, including CART decision trees and neural networks, and find that decision trees were able to correctly classify whether data was from the non-interference scenario or the interference scenario 75.86% of the time for TCP traffic and 91.43% of the time for UDP traffic.

Year of Submission

2022

Department

Department of Computer Science

University Honors Designation

A thesis submitted in partial fulfillment of the requirements for the designation University Honors

Date Original

5-2022

Object Description

1 PDF file (vi, 39 pages)

Language

en

File Format

application/pdf

Download
Related Work

Comments?

Share

COinS