Security and Resilience of Multi-Bitrate, Low-Power Lossy IoT Networks
Tez Türü: Bütünleşik Doktora
Tezin Yürütüldüğü Kurum: University of Bristol, Engineering Faculty, Electrical and Electronic Engineering , İngiltere
Tez Danışmanı: Dr. George Oikonomou
Tezin Onay Tarihi: 2023
Tezin Dili: İngilizce
Desteklendiği Program: Diğer
Özet:
L
ow power Lossy Internet of Things (IoT) systems have gained the attention of researchers
and industry in recent years. The overall research on resilient IoT networks aims to address connectivity issues, enhance data collection mechanisms, fortify security against
vulnerabilities at both the physical layer and in data collection practices, and develop robust
countermeasures against potential threats. Recognising the limitations of these severely constrained battery-powered IoT devices, such as restricted power and memory, the research aims to
contribute practical insights for deploying resilient solutions in low-power IoT networks across
diverse applications.
In this thesis, a novel multi-band multi-radio concept is demonstrated. This concept, along
with the resilient IoT standards of Time Slotted Channel Hopping (TSCH), Routing Protocol for
Low-Power and Lossy Networks (RPL) routing, and Constrained Application Protocol (CoAP)
protocols, is shown to improve network connectivity and resiliency. The most notable connectivity
improvement was found in the increased range of communications and bandwidth as a result
of the multi-bit rate, multi-frequency radios utilised in a hybrid manner within the network. It
has also demonstrated that improved bandwidth in the network and reduced hop counts give the
ability to improve packet delivery rate while not increasing energy consumption.
To provide more resilience, anomalous data was injected into a real-world dataset in the form
of a function of the original data to make detection difficult. The machine learning algorithms
detected the anomalies in the dataset, proving that they can be used to detect and filter drifts on
the sensor readings.
Finally, a lightweight implementation of secret key generation and establishment is also
investigated regarding processing power and energy consumption. The results show that the key
transport protocol surpasses the software implementation of Elliptic Curve Cryptography (ECC)
by hundreds of times in terms of energy for 2 different devices used.