Self-Reliance for the Internet of Things: Blockchains and Deep Learning on Low-Power IoT Devices
Doktorsavhandling, 2022

The rise of the Internet of Things (IoT) has transformed common embedded devices from isolated objects to interconnected devices, allowing multiple applications for smart cities, smart logistics, and digital health, to name but a few. These Internet-enabled embedded devices have sensors and actuators interacting in the real world. The IoT interactions produce an enormous amount of data typically stored on cloud services due to the resource limitations of IoT devices. These limitations have made IoT applications highly dependent on cloud services. However, cloud services face several challenges, especially in terms of communication, energy, scalability, and transparency regarding their information storage. In this thesis, we study how to enable the next generation of IoT systems with transaction automation and machine learning capabilities with a reduced reliance on cloud communication. To achieve this, we look into architectures and algorithms for data provenance, automation, and machine learning that are conventionally running on powerful high-end devices. We redesign and tailor these architectures and algorithms to low-power IoT, balancing the computational, energy, and memory requirements.

The thesis is divided into three parts:
Part I presents an overview of the thesis and states four research questions addressed in later chapters.
Part II investigates and demonstrates the feasibility of data provenance and transaction automation with blockchains and smart contracts on IoT devices.
Part III investigates and demonstrates the feasibility of deep learning on low-power IoT devices.
We provide experimental results for all high-level proposed architectures and methods. Our results show that algorithms of high-end cloud nodes can be tailored to IoT devices, and we quantify the main trade-offs in terms of memory, computation, and energy consumption.

Internet of Things

Deep Neural Networks

Smart Contracts



Lecture hall: EE - 6233 , EDIT Building 6th floor, Hörsalsvägen 11, Campus Johanneberg.
Opponent: Associate Professor (Docent) Shahid Raza at Uppsala University & Director of Cybersecurity Unit at RISE Research Institutes of Sweden,


Christos Profentzas

Chalmers, Data- och informationsteknik, Nätverk och system

IoTLogBlock: Recording Off-line Transactions of Low-Power IoT Devices Using a Blockchain

Proceedings of the 44th IEEE Conference on Local Computer Networks (LCN),; (2019)

Paper i proceeding

TinyEVM: Off-Chain Smart Contracts on Low-Power IoT Devices

40th IEEE International Conference on Distributed Computing Systems,; (2020)p. 507-518

Paper i proceeding

Performance of Secure Boot in Embedded Systems

Proceedings of the 15th IEEE International Conference on Distributed Computing in Sensor Systems (DCOSS),; (2019)p. 198-204

Paper i proceeding

Performance of deep neural networks on low-power IoT devices

CPS-IoTBench '21: Proceedings of the Workshop on Benchmarking Cyber-Physical Systems and Internet of Things,; (2021)

Paper i proceeding

Christos Profentzas, Magnus Almgren, Olaf Landsiedel, “MicroTL: Transfer Learning on Low-Power IoT Devices"

Christos Profentzas, Magnus Almgren, Olaf Landsiedel, “MiniLearn: On-Device Learning for Low-Power IoT Devices"

In the early days of the Internet, the communication was done typically through a telephone line with notorious long cables and extra hardware (remember the 56k modem?). Nowadays, connecting to the Internet has become simple for daily users, especially through wireless communication directly integrated into smart devices. Even though the capabilities of our common devices have increased, being smart has nothing to do with super capabilities, nor having a supercomputer on every device in our houses, but all a smart device has to do is communicate with a large network of computers with more resources that provide more services than a single computer can do. What may not be apparent to daily users is that wireless communication is quite energy-hungry, and smart devices have a strict energy budget running sophisticated algorithms to balance their limited resources.
        Nowadays, we use the term cloud to describe the cluster of computers that appear to offer an abundance of memory and computational capabilities using the Internet. In this regard, the collection of smart devices with sensors and actuators that connect with the cloud is called the Internet of Things (IoT), and this thesis studies the small ones like the smartwatch some of us may wear right now while the cloud analyzes our health & performance. Even though the cloud provides intelligent services, it sacrifices users' privacy (does anyone read the "terms and conditions"?). Moreover, IoT applications need to store multiple interactions and agreements among devices and cloud services distributed in diverse geographical locations, owned or operated by different parties. For example, resolving the parking fee for a rental electric scooter in the city center will need several interactions with the cloud.
        Even though the cloud appears to be unlimited, IoT devices produce data at a pace that raises questions if the cloud can handle them. Moreover, IoT reliance on cloud services gives access to important personal information. This thesis explores questions about the extent to which our smartwatch can rely on its own effort and capabilities to provide intelligent services to us with reduced cloud communication and preserve our privacy. We answer by redesigning algorithms that typically run on powerful devices to small devices balancing the computational, energy, and memory resources.

RIOT: Ett resilient sakernas internet

Myndigheten för samhällsskydd och beredskap (MSB2018-12526), 2019-01-01 -- 2023-12-31.

KIDSAM: Kunskap- och informationssdelning i digitala samverkansprojekt

VINNOVA (2018-03966), 2018-11-01 -- 2021-11-30.

AgreeOnIT: Lättvikts konsensus och distribuerat datakunskap i resursbegränsade sakernas Internet

Vetenskapsrådet (VR) (37200024), 2019-01-01 -- 2022-12-31.



Datavetenskap (datalogi)


Informations- och kommunikationsteknik



Doktorsavhandlingar vid Chalmers tekniska högskola. Ny serie: 5131



Lecture hall: EE - 6233 , EDIT Building 6th floor, Hörsalsvägen 11, Campus Johanneberg.


Opponent: Associate Professor (Docent) Shahid Raza at Uppsala University & Director of Cybersecurity Unit at RISE Research Institutes of Sweden,

Mer information

Senast uppdaterat