Energy-neutral Event Monitoring for Internet of Nano Things
Thesis submitted in fulfilment of the requirements for the degree of
Doctor of Philosophy in Computer Science and Engineering
Najm Hassan
Supervisors: Prof. Mahbub Hassan and Prof. Chun Tung Chou
November 2018 PLEASE TYPE THE UNIVERSITY OF NEW SOUTH WALES Thesis/Dissertation Sheet
Surname or Family name: Hassan
First name: Najm Other name/s:
Abbreviation for degree as given in the University calendar: PhD
School: Computer Science and Engineering Faculty: Engineering
Title: Energy-neutral Event Monitoring for Internet of Nano Things
Abstract 350 words maximum: (PLEASE TYPE)
Advancements in nanotechnology promise new capabilities for the Internet of Things (IoT) to monitor extremely fine-grained events with sensors as small as up to a hundred nanometers. Researchers predict that such tiny sensors can be connected to the Internet using graphene-based nano-antenna radiating in the terahertz band, giving rise to the so called Internet of Nano-Things (IoNT). Powering such wireless communications with nanoscale energy supply, however, is a major challenge to overcome. Since in many application domains, different types of events discharge different amounts of energy to the environment, we propose an energy-neutral event monitoring framework, called eNEUTRAL IoNT, that allows the sensors to transmit event information using only the amount of energy harvested from the events. We design and analyse two implementation methods for this framework. The first method uses a single pulse containing the entire energy harvested from the event but manipulates its pulse width (time duration) to create unique pulse amplitude for a given combination of event type and its location. In the second option, the harvested event energy is divided into two pulses so that the energy of the first pulse uniquely defines a location and the second pulse uses the remaining energy to identify event types. To minimize classification error at the receiver, we optimize pulse durations in the single-pulse option and pulse energies in the dual-pulse option. Feasibility of eNEUTRAL IoNT is demonstrated using extensive numerical experiments involving terahertz channels. We find that the dual-pulse approach significantly outperforms the single-pulse approach achieving 99% accuracy for detecting both location and event type in 10-node network monitoring two different event types for a radius of 28 mm. As nanoscale energy harvesters and transmitters are still not available to realize operational event monitoring nodes, we, therefore, evaluate and design a key component of the IoNT system, namely pulse generator, using COMSOL Multiphysics. We first surveyed the literature of different approaches for pulse generation that generate Surface Plasmon Polaritons (SPPs) which lead to femtosecond long pulses in graphene. Based on our analysis, we found that most of the existing configurations require complex structures including a prism, periodic slits and special circuits which may be difficult to implement in severely resource constraint nodes. Using COMSOL Multiphysics, we show that pulses in the terahertz band can be generated using the near-field method by which the matching condition for excitation of SPPs can be easily satisfied. The performance of the proposed near-field excitation method in the terahertz band is studied via numerical simulations. This includes the choice of the source, its phase angle, the chemical potential, and frequency. The proposed model can be a good candidate for a low-complexity realization of a THz pulse generator in tiny IoNT nodes. We believe that our findings will open the door for a new direction of research and development toward the energy-neutral event monitoring systems in IoNT.
Declaration rel ating to disposition of project thesis/dissertation
I hereby grant to the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or in part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all property rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation.
I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstracts International (this is applicable to doctoral theses only).
…………………………………………………………… ……………………………………..……………… ……….……………………...…….… Signature Witness Signature Date
The University recognises that there may be exceptional circumstances requiring restrictions on copying or conditions on use. Requests for restriction for a period of up to 2 years must be made in writing. Requests for a longer period of restriction may be considered in exceptional circumstances and require the approval of the Dean of Graduate Research.
FOR OFFICE USE ONLY Date of completion of requirements for Award:
ORIGINALITY STATEMENT
I hereby declare that this submission is my own work and to the best of my knowl- edge it contains no materials previously published or written by another person, or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgment is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project’s design and conception or in style, presentation, and linguistic expression is acknowledged.
Signed ...... Najm Hassan November 2018 COPYRIGHT STATEMENT
I hereby grant the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or part in the University libraries in all forms of media, now or hereafter known, subject to the provisions of the Copyright Act 1968. I retain all proprietary rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation. I also authorise University Micro- films to use the 350 word abstract of my thesis in Dissertation Abstract Interna- tional (this is applicable to doctoral theses only). I have either used no substantial portions of copyright material in my thesis, or I have obtained permission to use copyright material; where permission has not been granted I have applied/will apply for a partial restriction of the digital copy of my thesis or dissertation.
Signed ...... Najm Hassan November 2018
AUTHENTICITY STATEMENT I certify that the Library deposit digital copy is a direct equivalent of the final officially approved version of my thesis. No emendation of content has occurred and if there are any minor variations in formatting, they are the result of the con- version to digital format.
Signed ...... Najm Hassan November 2018 i
ABSTRACT Advancements in nanotechnology promise new capabilities for the Internet of Things (IoT) to monitor extremely fine-grained events with sensors as small as up to a hundred nanometers. Researchers predict that such tiny sensors can be connected to the Internet using graphene-based nano-antenna radiating in the terahertz band, giving rise to the so called Internet of Nano-Things (IoNT). Pow- ering such wireless communications with nanoscale energy supply, however, is a major challenge to overcome. Since in many application domains, different types of events discharge different amounts of energy to the environment, we propose an energy-neutral event monitoring framework, called eNEUTRAL IoNT, that al- lows the sensors to transmit event information using only the amount of energy harvested from the events. We design and analyse two implementation methods for this framework. The first method uses a single pulse containing the entire en- ergy harvested from the event but manipulates its pulse width (time duration) to create unique pulse amplitude for a given combination of event type and its loca- tion. In the second option, the harvested event energy is divided into two pulses so that the energy of the first pulse uniquely defines a location and the second pulse uses the remaining energy to identify event types. To minimize classifica- tion error at the receiver, we optimize pulse durations in the single-pulse option and pulse energies in the dual-pulse option. Feasibility of eNEUTRAL IoNT is demonstrated using extensive numerical experiments involving terahertz chan- nels. We find that the dual-pulse approach significantly outperforms the single- pulse approach achieving 99% accuracy for detecting both location and event type in 10-node network monitoring two different event types for a radius of 28 mm. As nanoscale energy harvesters and transmitters are still not available to realize operational event monitoring nodes, we, therefore, evaluate and design a key component of the IoNT system, namely pulse generator, using COMSOL ii
Multiphysics. We first surveyed the literature of different approaches for pulse generation that generate Surface Plasmon Polaritons (SPPs) which lead to fem- tosecond long pulses in graphene. Based on our analysis, we found that most of the existing configurations require complex structures including a prism, pe- riodic slits and special circuits which may be difficult to implement in severely resource constraint nodes. Using COMSOL Multiphysics, we show that pulses in the terahertz band can be generated using the near-field method by which the matching condition for excitation of SPPs can be easily satisfied. The per- formance of the proposed near-field excitation method in the terahertz band is studied via numerical simulations. This includes the choice of the source, its phase angle, the chemical potential, and frequency. The proposed model can be a good candidate for a low-complexity realization of a THz pulse generator in tiny IoNT nodes. This thesis is dedicated to my beloved wife for her love, support and encouragement, To my lovely little son Ammar Hassan and To Ammar’s grandparents for their timeless love iv
ACKNOWLEDGEMENTS
Working at the School of Computer Science and Engineering Computer Sci- ence and Engineering School (CSE) at the University of New South Wales University of New South Wales (UNSW) has been a great pleasure and an incredible privi- lege. First, I would like to express my sincere appreciation and profound gratitude to my supervisor, Professor Mahbub Hassan for his exceptional support, encour- agement, and guidance during all stages of this research. His truly valuable aca- demic excellence, scientific intuition, and a beautiful mind have made him a con- stant oasis of ideas and passions in science. This has inspired and enriched my growth as a student. I also express my genuine thanks to Professor Chun Tung Chou, as my joint-supervisor, for his kind help, incredible support and valuable discussion. It was an honor for me to work closely with such a talented, polite and creative personality. In addition, I express my sincere appreciation to Dr. Ming Ding for great discussion and support. It was an excellent opportunity for me to have him in my supervision panel. I also express my genuine thanks to our external collaborator Dr. Marios Mattheakis, Harvard University USA, for his incredible support contributing to my thesis. Most of all, I am profoundly and forever indebted to my parents for their never-ending love, and encouragement throughout my entire life. Unfortu- nately, I lost my mother in the third year of my PhD. It was a tragedy for me as she was my soul mate. I am sure my parents’ souls is always praying for me and I also continuously pray for the repose of their souls. Last but no means least, it gives me immense pleasure to thank Dr. Eisa Zare- pour and lab members I worked with in CSE UNSW. I enjoyed my PhD study because of them. I had a great life and study experience. They are wonderful v people and always ready to help. Finally, I would like to thank my family for their unwavering understanding and supports. I thank them mainly for their understanding and the acceptance that I have been so busy during my PhD study that we have hardly been able to spend much time together. They are my source of strength, and without their tremendous support, this thesis would never have been started.
Najm Hassan Sydney, Australia November 2018 Contents
1 Introduction1 1.1 Motivation...... 3 1.2 Problem Statement...... 5 1.3 Contributions...... 6 1.4 Dissertation Organization...... 10
2 Background and the State of the Art 12 2.1 Internet of Nano-Things...... 13 2.2 Building blocks of IoNTs...... 13 2.2.1 Nanosensor Node...... 14 2.2.2 Propagation Channel...... 19 2.2.3 Receiver...... 22 2.3 Applications...... 25 2.3.1 Medical Applications...... 25 2.3.2 Industrial Applications...... 28 2.3.3 Environmental Applications...... 28 2.4 Sensor powering challenges...... 29 2.5 Pulse generation for tiny IoNT nodes...... 31
3 Energy-neutral Single Pulse Transmission 38 3.1 Introduction...... 39 3.2 System Model...... 40
vi CONTENTS vii
3.2.1 Event Model...... 42 3.2.2 Structure of the monitoring node...... 43 3.2.3 Pulse Transmitter Model...... 44 3.2.4 Channel Model...... 46 3.2.5 Pulse Receiver Model...... 46 3.3 Single Pulse Transmission (SPT)...... 48 3.3.1 Problem description...... 48 3.3.2 Pulse Width Allocation (PWA)...... 52 3.3.3 Numerical investigation on PWA...... 54 3.3.4 Impact of pulse widths on error probability...... 54 3.3.5 Pulse widths allocation to more than two nodes...... 58 3.4 Evaluation...... 59 3.4.1 Impact of node-RS distance...... 60 3.4.2 Impact of number of nodes...... 60 3.4.3 Impact of number of event types...... 61 3.5 Chapter Summary...... 62
4 Energy-neutral Dual Pulse Transmission 64 4.1 Introduction...... 64 4.2 Dual Pulse Transmission (DPT)...... 65 4.2.1 Problem description...... 66 4.2.2 Pulse Energy Detection...... 67 4.3 Pulse Energy Allocation (PEA)...... 68 4.4 Performance Evaluation...... 69 4.4.1 Impact of the network radius...... 70 4.4.2 Impact of the number of nodes...... 72 4.4.3 Impact of the number of event types...... 72 4.4.4 Impact of the event energy spread...... 73 4.4.5 Impact of the node placement...... 75 CONTENTS viii
4.4.6 Impact of the event energy fluctuations...... 76 4.5 Discussion...... 77 4.6 Chapter Summary...... 78
5 Pulse Generator for tiny IoNT Nodes 79 5.1 Introduction...... 80 5.2 SPP Preliminaries...... 81 5.3 Discussion...... 83 5.4 Pulse Generation for Event Monitoring Node...... 84 5.5 Simulation and Numerical Results...... 86 5.5.1 Impact of the frequency on the amplitude of SPP...... 87 5.5.2 Impact of the chemical potential on the SPP resonance... 87 5.5.3 Impact of the phase angle of the evanescent source on SPP resonance...... 90 5.5.4 Impact of the type of evanescent source on the SPP resonance 90 5.6 Chapter Summary...... 91
6 Conclusion and Future Works 93 6.1 Concluding Remarks...... 93 6.2 Future Directions...... 95
Bibliography 97
A Acronyms 113
Appendix 113 List of Figures
2.1 An integrated nanosensor node [1]...... 14 2.2 Interface of the HITRAN on the web tool. Molecular absorption coefficient can be calculated directly from HITRAN database.... 21 2.3 Total path-loss in dB of THz band at different distances...... 22 2.4 Molecular Absorption Noise PSD in dB/Hz at different distances in THz band...... 23 2.5 Block diagram of energy based detection where the pulse energy is detected by integrating the received energy over an Integration
window (Tint)...... 23
2.6 In Energy detection, the integration window (Tint) is greater than the pulse width (duration) so the additional noise is also integrated. 24 2.7 Block diagram of a CTMA based detection to detect either peak power (pulse amplitude) or pulse energy...... 24 2.8 Network architecture for the IoNT: (a) human lungs monitoring [2] (b) monitoring chemical reactions [3] and (c) plant monitoring [4]. 26 2.9 Generating SPPs using different excitation methods...... 32 2.10 Schematic representation of Attenuated Total Reflection (ATR) meth- ods which are used for SPP excitation: (a) Kretschmann -Raether and (b) Otto Configuration...... 34
3.1 Illustration of eNEUTRAL event monitoring node which includes an energy harvester and a radio...... 43
ix List of Figures x
3.2 CTMA detector to detect the peak power of the pulse...... 47 3.3 Block diagram of the event monitoring system...... 49 3.4 Illustration of pulse widths vs peak power of transmitted pulses in single-pulse approach...... 51 3.5 (a) Error probability versus pulse width of two nodes at distance d = 30 mm (b) A cross-section of heatmap...... 57 3.6 Energy ratio vs pulse width ratio: (a) Local minimum (b) Global minimum...... 58 3.7 The effect of distance on the classification error in single-pulse ap- proach for 5-, 10- and 20-node cases...... 60 3.8 Effect of number of nodes on the classification error for a 30mm network monitoring two distinct event types with energies 1 aJ and 2 aJ...... 61 3.9 Effect of number of event types on classification error for a 5-node network of radius 30 mm...... 62
4.1 CTMA detector to detect the pulse energy...... 68 4.2 The effect of distance on the classification error in SPT and DPT for 5- and 10-node cases...... 71 4.3 Effect of number of nodes on achievable radius for target error be- 2 3 tween 10− and 10− when monitoring two distinct event types with energies 1 aJ and 2 aJ...... 72 4.4 Effect of number of nodes on the classification error for a 30 mm network monitoring two distinct event types with energies 1 aJ and 2 aJ...... 73 4.5 Effect of number of event types on classification error for a 5-node network of radius 30 mm...... 74 4.6 Effect of number of events types on achievable radius for target 2 3 error between 10− and 10− for a 5-node network...... 74 List of Figures xi
4.7 Effect of placement error for a 5-node network with a radius of 30 mm for 2 distinct event types having energies of 1 aJ and 1.5 aJ.. 76 4.8 Effect of event energy fluctuation for a 5-node network with a ra- dius of 30 mm for 2 distinct event types...... 77
5.1 Illustration of a pulse generator where SPP is excited on graphene sheet using the near-field method. A nanoantenna takes the SPP wave and radiates it into a free-space EM wave...... 85 5.2 The magnetic intensity I demonstrates the plasmon resonance on a doped graphene monolayer (dotted white line). The spatial distri- bution of I¯ is represented in arbitrary values (a.u.) by (a) colorbar in x-y plane and by (b) with a solid blue line along the graphene layer (y = 0). An evanescent TM EM dipole source (n=1) of fre- quency f = 10 THz and with phase angle θ = π/2 is located at
Xs = 100 µm and Ys = 30 nm above the graphene layer, indicated by a tiny white spot. The chemical potential of graphene is con-
sidered µc = 0.5 eV. The SPP wavelength is calculated to be λsp = 5 µm...... 88 5.3 The maximum value of I¯ for several values of frequency is pre- sented revealing that the higher the frequency the better plasmon resonance is achieved...... 89
5.4 The maximum value of I¯ for several values of doping µc is shown revealing that at a certain frequency of f = 10 THz, the plasmon
resonance is a increasing function of the doping for values µc < 0.4 eV...... 89 5.5 The maximum value of I¯ for several as function as the phase an- gle θ of the source is demonstrated with blue solid line. The best plasmon resonance is observed for the θ = π/2...... 90 List of Figures xii
5.6 The plasmon profile intensity I¯ along the graphene sheet is repre- sented for several values of evanescent parameter n of the source, that is, for n = 1, n = 2 and n = 3...... 91 List of Tables
2.1 Power consumption of different types of nanosensors versus power harvested at nanoscale...... 19
3.1 Table of the most frequently used parameters in single-pulse ap- proach of eNEUTRAL IoNT framework...... 53 3.2 The Composition of normal air...... 54 3.3 Pulse widths and the error probability for 2-5 nodes at distance d = 30 mm...... 59
4.1 Comparison between SPT and DPT...... 66 4.2 Table of the most frequently used parameters in dual-pulse solu- tion of eNEUTRAL IoNT framework...... 70 4.3 Impact of energy gap of different event types on the error proba- bility at distance d = 30 mm...... 75
xiii List of Publications xiv
List of Publications
Journal Articles
• Najm Hassan, Chun Tung Chou, Mahbub Hassan,“eNEUTRAL IoNT: Energy- neutral Event Monitoring for Internet of Nano Things,” under revision in IEEE Internet of Things Journal (IoT-J), 2018.
• Najm Hassan, Marios Mattheakis, Ming Ding, “Sensorless Node Architecture for Events Detection in Self-Powered Nanosensor Networks,” Nano Communi- cation Networks (Elsevier) Journal, vol. 19, pp 1-9, November 2018.
Conference Proceedings
• N Hassan, C. T. Chou, M Hassan, “Event and Node Identification from a Single-Pulse Transmission in Self-powered Nanosensor Networks,” in Proceed- ings of 4th ACM International Conference on Nanoscale Computing and Communication (NANOCOM), September 27-29, 2017, Washington D.C., DC, USA.
• E Zarepour, N Hassan, M Hassan, C. T. Chou, and M E Warkiani, “Design and Analysis of a Wireless Nanosensor Network for Monitoring Human Lung Cells,” in Proceedings of 10th EAI International Conference on Body Area Networks (BodyNets), 2015, September 28-30, Sydney, Australia. Chapter 1
Introduction
The last decade has witnessed significant advances in nanotechnology which makes it possible to fabricate sensor nodes at nanoscale at or below a hundred nanometers. These nodes are made from novel materials which have unique physical, electrical and optical prosperities. Such nodes have the capabilities to sense molecule level events in immediate surroundings [1]. Recent studies have revealed that these nanosensors can communicate using graphene-based nano- antenna radiating in the terahertz band (0.1-10 THz) [5,6] ushering new Internet of Thing (IoT) capabilities for gathering knowledge at an unprecedented depth and scale. Researcher are now pursuing this new direction of IoT under the ban- ner of Internet of Nano Thing (IoNT)[7] with nanoscale monitoring techniques explored for human body [2,8], plants [4], chemical processes [9], and so on. There is a large volume of published studies related to Wireless Sensor Net- work (WSN), however, communication at the nanoscale is relatively a new re- search topic. In recent years, the area of nanocommunication has significantly attracted the research community mainly due to its unprecedented applicability in different applications. To name a few, biomedical, industrial, environmental and military are envisaged applications where nanoscale communication is ex- pected to be deployed to empower bottom-up control of molecule level events. In biomedical health monitoring systems, nanosensor networks can be deployed
1 Chapter 1 Introduction 2 to monitor cholesterol, sodium, glucose ions in blood or to monitor different infectious agents [1]. There are some reported applications of nanoscale com- munication in industrial such as real-time monitoring and controlling chemical processes at the molecular level [2, 10]. In environmental applications, plant monitoring systems, plagues defeating systems are complementary applications of nanoscale communication [1]. On the military and defense side, the nanoscale communication applications are nuclear, biological and chemical (NBC) defenses, where chemical and biological nanosensors can be used to detect harmful chem- icals and biological arsenal in the distributed manner. The other military ap- plications are damage detection systems which can be used to detect very small cracks in textiles, vehicles and rockets etc [1]. In such applications, sensed in- formation and measurements can be transmitted to a distinct receiver by estab- lishing nanoscale wireless communication within the IoNT paradigm. How to balance energy expenditure so that it never exceeds the harvested energy in tiny IoNT nodes is a challenging problem to solve. Therefore, novel sensorless energy- neutral solutions are required to monitor events in the environment without em- ploying onboard sensors. Such solutions can obviate the need for several con- ventional sensor node elements (i.e. microprocessor, memory, and the sensor) to avoid the problem of powering, however, it can also limit the sensing capability locally. Instead, the molecule level events sensing and monitoring can be accom- plished remotely at a distant receiver which is least explored. If different types of events emit different amounts of energy to the environment, then technically it is feasible to identify events, from their emitted energy, at the receiver. In ad- dition to events detection, one of the major issues is to determine the identity of nodes in severely resource constraint nanoscale communication. In the cur- rent IoT system, the packet-based conventional addressing scheme is used where a large number of bits within the payload, header, and the preamble, for each packet, are transmitted. Adopting a packet-based approach of IoT system will require much power to transmit many bits which can be energy inefficient in Chapter 1 Introduction 3 severely resource constraint IoNT systems. The aim of this thesis is to propose novel packet-less energy-neutral solutions to overcome the power consumption limitation. An energy-neutral solution aims to balance the energy expenditure in IoNT nodes so that it never exceeds the harvested energy. In this direction, we propose energy-neutral event monitoring schemes tailored with the limited energy and restricted resources of tiny sen- sors. At the nanoscale, energy harvesters and transmitters are still not available to realize operational IoNT system. Therefore, this thesis will also model the key component of the tiny IoNT node, i.e., a pulse generator, using COMSOL Multi- physics. The proposed pulse generator is simple to construct and uses a near-field method to generate a pulse in the THz band. We expect that simple structures are more likely to succeed in the future design of severely resource constraint nodes.
1.1 Motivation
Recent studies confirm that nanosensors may be able to communicate in the Terahertz (THz) band as the transmission band, giving rise to the so-called IoNTs [5–7]. As THz band is the resonance frequency of molecules, communication in this band is severely affected by attenuation and molecular absorption noise [10, 11]. Although, there has been some recent work in the literature to ad- dress these issues. Recently, novel communication elements, channel modeling, and modulation schemes, and network architectures have been investigated for IoNTs [2,6, 10, 12–15]. In this direction, graphene-based THz nanoantenna and nanotransceiver which can generate THz pulses suitable for nanoscale communi- cation are investigated in [6,13]. Novel communication protocols and THz chan- nel models are developed [2,10,12]. To enable communication between nanosen- sors, novel network architectures are investigated in [2, 14, 15]. Recently, novel information coding techniques [16, 17] and energy harvesting schemes [18] are also proposed. Amongst all, pulse-based communication which is based on the Chapter 1 Introduction 4 transmission of hundred femtoseconds long pulses is proposed in [19]. Power supply at the nanoscale is a challenging problem. As batteries are dif- ficult to be built into and replaced in nanosystems, researchers are developing nanoscale energy harvesters (a.k.a nanogenerators) [20–25], that can scavenge tiny amounts of energy from the ambient environment to power event detection and very basic wireless communication where all data is transmitted as a series of extremely short (a few hundred femtoseconds) pulses [19]. The harvested energy can be used to transmit short pulses which can reduce the total energy consump- tion drastically compared to conventional continuous wave wireless communica- tions. However, as each pulse consumes a finite amount of energy, the harvested energy may not be adequate to transmit a large number of pulses uploading event information and sensor node identification each time events are detected in the environment. Interestingly enough, the information in the pulse based communication can be encoded either in the pulse with (duration), pulse amplitude or pulse position of the transmitted pulse. Therefore it is possible to use a pulse to convey address and event information using any of these parameters of the transmitted pulse. However, pulse position modulation requires synchronization between a trans- mitting node and a receiver which is difficult to achieve in resource-constrained nanoscale communication. We, therefore, in this thesis, only use pulse width (duration) and transmitted energy of a pulse to convey two pieces (address and event) of information. Another example of using the property of the transmitted signal to identify the transmitter can be found in [26] where the amplitude of the signal is used to identify the transmitter. However the focus of [26] is on medium access control in bacterial communication networks but our focus is on delivering node identity and event information in THz electromagnetic communication. Chapter 1 Introduction 5
1.2 Problem Statement
Powering wireless communication at the nanoscale is a major challenge. Be- cause, nanoscale energy harvesters (a.k.a nanogenerators) [20–25] can not gen- erate enough power to transmit a large number of pulses uploading event infor- mation and sensor node identification each time events are detected in the envi- ronment. Likewise, the insufficient harvesting rate in embedded environments may not allow continuous operation of sensors [27]. This means that there is a gap between the energy harvesting and the energy consumption which makes it hard to achieve self-powered IoNT systems. How to balance the use of energy so that it never exceeds the harvested energy in tiny IoNT nodes is a challenging problem to solve. Dealing with the finite amount of energy harvested at the nanoscale, the aim of this thesis is to design, develop and evaluate energy-neutral solutions for resource-constrained nanonetworks such as:
• To propose energy-neutral event monitoring solutions that allow the sen- sors to transmit event information (event type and its location) using only the amount of energy harvested from the events.
• To design pulses that will accurately convey both event type and its location to a distant receiver.
• To develop optimization models for pulse duration and pulse energy with the aim to improve the detection accuracy by minimizing the decoding er- ror at the receiver
• The last contribution in this thesis is to design and evaluate the key compo- nent of an IoNT system, i.e., pulse generator, using COMSOL Multiphysics which is simulation software close to the physical reality. Chapter 1 Introduction 6
1.3 Contributions
Using the properties of the transmitted signal to identify node and event types remotely, this thesis conducts a numerical study of the events monitoring associ- ated with resource-constrained nodes in self-powered nanonetworks within the IoNTs paradigm. Before highlighting the key contributions, we first start with system model as follows:
• We started with event model to identify event types which can be detected using the energy released by the events. An example application of the energy-neutral is to monitor the events within a chemical production pro- cess [11, 27] and demonstrate the feasibility of using emitted energy as the signature for the event. The key idea is to place a monitoring node at a site to monitor the amount of energy released by the chemical reactions taking place at the site. If different chemical reactions release the differ- ent amount of energy, then, technically, it is possible to use the amount of energy released by a chemical reaction as a signature to distinguish differ- ent reactions. However, in order to extend the concept to multiple nodes, we need to ensure that each node is only monitoring the energy released locally at the site and is minimally impacted by the energy released at its neighbouring sites. This can be realized if the sites are sufficiently far apart from each other.
• We then detailed the structure of monitoring node which consists of two components: an energy harvester and a radio (pulse generator). The pur- pose of the energy harvester is to convert the energy emitted by the event into electrical energy. We found that there are nanoscale energy harvesters [20–25, 28, 29] that can scavenge tiny amounts of energy from the events in the ambient environment. As such, an energy harvester that harvests the emitted energy from the event can also serve as a sensor for monitoring such events Chapter 1 Introduction 7
• Next, we evaluated pulse transmitter model for monitoring node to gener- ate a pulse using the harvested energy from an event. The purpose of the transmitter is to convert the electrical energy into a radio message. When an event occurs, the energy emitted by the event is harvested by suitable energy harvester and input to the pulse generator to generate a pulse.
• To enable communication among nanosensors, we have a communication channel between a node and a remote station where the transmitted pulse is attenuated. Recent studies confirm that these sensors may be able to communicate in the THz band [6, 30]. We, therefore, reviewed widely used THz channel model for nanonetworks within the IoNTs paradigm.
• Events classification is performed at the receiver. We, therefore, provided a detailed study of pulse detection models. We found that Continuous Time Moving Average (CTMA) detector can be used to detect peak power or pulse energy of received pulse [31].
• After building system model, we propose two energy-neutral event moni- toring solutions that allow the sensors to transmit event information using only the amount of energy harvested from the events. We then design, eval- uate these two implementation solutions for Energy-neutral (eNEUTRAL) IoNT. We also studied the decoding performances of both implementation options of eNEUTRAL IoNT.
The main contributions along with detail descriptions are summarized as fol- lows:
• Single Pulse Transmission (SPT): SPT is the first implementation solution of the eNEUTRAL IoNT that uses a single pulse containing the entire energy harvested from the event. However, it manipulates its pulse width (duration) to create a unique pulse amplitude for a given combination of event type and Chapter 1 Introduction 8
its location. This way SPT or simply single-pulse approach enables multi- ple nodes to communicate event type and its location to a distinct receiver by transmitting a single wireless pulse. However, single-pulse approach imposes the requirement that each node uses a particular pulse width (du- ration). Therefore, a major challenge is how to allocate pulse widths to multiple nodes to minimize the classification error. This way, we develop optimization model for pulse width allocation with the aim to minimize the classification error. We also study the relation between energy emitted by the events and the choice of pulse widths. From numerical simulation results, We find that single-pulse approach achieving 99% event type and location detection accuracy in 10-node network monitoring two different event types for a distance of 22 mm.
• Dual Pulse Transmission (DPT): In SPT, we manipulate the pulse width to create a unique pulse amplitude for a given combination of event type and its location. This results in the increasing number of classes to clas- sify which is a major variable in the error probability. if we break down the number of classes into smaller class with two pulses then we can im- prove classification accuracy. We, therefore, propose Dual Pulse Transmis- sion (DPT) in which the nodes will have the same architecture to maintain the sensorless architecture but use a different mechanism for encoding and decoding. In the DPT or dual-pulse approach, the harvested event energy is divided into two pulses so that the energy of the first pulse uniquely de- fines a location and the second pulse uses the remaining energy to identify event types. We, therefore, encode the address and event type in the energy level of the transmitted pulse. We develop an optimization model for pulse energy allocation to optimize pulse energies. This optimal policy aims to improve the detection accuracy by minimizing classification error at the re- ceiver. Chapter 1 Introduction 9
• Using extensive numerical simulation we evaluate the performances of both single and dual pulse transmissions. The parameters that are considered to evaluate the performances of both systems are the distance, number of nodes, node placement, number of event types, event energy spread and stochastic energy harvesting (energy fluctuations) of different events. We find that the dual-pulse approach significantly outperforms the single- pulse approach achieving 99% accuracy for detecting both location and event type in 10-node network monitoring two different event types for a radius of 30 mm.
• Pulse generation for tiny IoNT nodes: Due to the severe volume restric- tions and low complexity of event monitoring nodes, the pulse generator should be kept simple and to generate THz signals without requiring com- plex structure. We, therefore, propose a graphene-based pulse generator to generate THz pulses. We studied different configurations which are used to generate SPPs on the metallic and graphene surface. The generated SPPs lead to femtoseconds long pulses in graphene. We note that the choice of the most suitable method depends on the frequency, the type of source, the ma- terial permittivity, and the system configuration. However, we find that our proposed generator can generate THz pulses directly on the graphene sur- face without adding any new circuit to event monitoring nodes. We model and evaluate the pulse generator in COMSOL Multiphysics which is close to the physical reality. We use near-field method to generate pulses in the THz band. We expect that simple structures are more likely to succeed in the future design of nano nodes. Chapter 1 Introduction 10
1.4 Dissertation Organization
The remainder of this thesis is organized as follows. In Chapter 2, we detail the state-of-the-art of the research related to this dissertation. We first overview the components of the network architecture of the Internet of Nano-Things. We then present the channel modeling and applications of nanosensor networks within the Internet of Nano-Things paradigm. Finally, we give separate related work sections for technical contributions so as to give the reader an appropriate back- ground respective of that particular chapter. In Chapter 3, we first introduced the system model, in details, for the eNEU- TRAL IoNT framework. We then propose the first implementation solution of the eNEUTRAL IoNT framework that uses a single pulse containing the entire energy harvested from the event. We motivate the need of numerical investiga- tion of pulse width allocation in two node scenario. We also study the impact of pulse widths on error probability. Finally, we investigate pulse width allocation and develop an optimization model to optimally allocate Pulse widths to multi- ple nodes with the aim to minimize the classification error. In Chapter 4, we introduce the second solution called dual-pulse transmission for the eNEUTRAL IoNT framework with the aim to improve the detection ac- curacy for longer distance. This approach uses a different encoding mechanism by encoding the address and event type in the energy level of the transmitted pulse. That is, the harvested event energy is divided into two pulses so that the energy of the first pulse uniquely defines a location and the second pulse uses the remaining energy to identify event types. To minimize classification error at the receiver, we develop an optimization model to optimize pulse energies and study the impact of address and event pulse energy on the classification error. In the end, we perform extensive numerical simulation and compare the perfor- mance and scalability of both solutions. We find that the dual-pulse approach significantly outperforms the single-pulse approach. Chapter 1 Introduction 11
In Chapter 5, we present a THz pulse generator based on near-field excita- tion for severely resource constraint nodes. We model the pulse generator, us- ing COMSOL Multiphysics. The proposed model is simple to construct without adding any new circuit and can generate pulses in the THz band. The thesis concludes in Chapter 6, followed by main future directions. Chapter 2
Background and the State of the Art
A nanosensor with limited sensing, restricted computational power, and highly constrained storage capabilities can sense, gather and share knowledge at the molecular level, empowering bottom-up control of many applications. It per- forms very simple computation, sensing and actuation tasks limited to close proximity. To increase their capabilities, these devices can be connected to per- form collaborative tasks in a distributed manner and to send the sensed data to the external world for analysis [2] under the banner of the Internet of Nano- Things (IoNTs). Nevertheless, as IoNT is in an early stage of development and has not been successfully demonstrated yet, most work is focused on mathematical modeling, and some conceptual models have been proposed recently. IoNT is a new challenging paradigm for researchers with known nano-specific challenges as discussed in Chapter 1. For this purpose, a literature survey is car- ried out to identify recent trends and development in the area of IoNT. In this Chapter, we present the discussions in such a way as to lead the readers toward the main contributions of the thesis. We will start with the introduction of IoNTs, followed by the background of the basic building blocks of IoNT system. We will then present a few proposed applications with related work for nanosensor net- works within the IoNT paradigm. Finally, we give separate related work sections for this thesis contributions so as to give the readers an appropriate background
12 Chapter 2 Background and the State of the Art 13 respective of that particular contribution.
2.1 Internet of Nano-Things
Internet of Nano-Thing is a new paradigm which connects nanosensors with ex- isting communication networks, and eventually, the Internet which leads to the development of next-generation standard based on Internet of Thing (IoT) called the Internet of Nano-Thing (IoNT) [7]. IoNT will open new doors of research in the area of nanosystems, and nanocommunication to gather knowledge at an un- precedented depth and scale [32]. This development is due to the recent advances in nanotechnology which has provided nanoscale sensing and monitoring solu- tions to many real-world applications like biotechnology and biomedical, agri- culture, and industry [2,4,8,9]. With the emergence of IoNT, researchers propose new communication stan- dards for nanoscale devices to communicate with each other in diverse appli- cations. IoNT uses two broad areas of nanoscale communication such as nano- electromagnetic and molecular communication. Nano-electromagnetic commu- nication is defined as the as transmission and receiving of Electro-Magnetic (EM) radiation among nanoscale devices whereas the molecular communication is re- garded as the transfer of information using molecules and is therefore referred to as molecular communication [7,32]. The focus of this thesis is on electromagnetic communication.
2.2 Building blocks of IoNTs
The Internet of Nano-Thing architecture comprises three main components: nanosen- sor node, propagation channel and a receiver. In this section, we present the literature for each of these components. Chapter 2 Background and the State of the Art 14
A sensor node senses and monitors molecule level events. These sensed mea- surements are processes locally and are transmitted via the communication chan- nel to a distinct receiver. A receiver is a macro scale device which receives the data for analysis. In what follows, we provide details about these components.
2.2.1 Nanosensor Node
In recent years, the nanotechnology advances made it possible to develop a new generation of smaller electronics up to a hundred nanometers. Similar to the macro-level sensor nodes, these nanoscale devices will have nanocomponents, including processing, sensing, transmitting, and power storage unit respectively as shown in Figure 2.1. Nano-processor Nano-transceiver Nano-antenna Nano-sensors Nano-memory Nano-actuator Nano-capacitor
Energy-harvester
Figure 2.1: An integrated nanosensor node [1].
Nano-processor
The nano-processor is the main component which drives all the onboard nano- electronics with the exception of the energy harvester and nanobatteries [33]. Due to the form factor of a sensor node, a nano-processor must contain an appro- priate number of transistors to accomplish diminutive tasks. For example, the first nano-processor called Nano-Sensor Data Processor (NSDP) with 4-bit data Chapter 2 Background and the State of the Art 15 processing is designed in [34]. This nano-processor is based on basic Processing Element (PE) which is a cell with 20 nm by 20 nm in size. Each cell is a square nanostructure with a quantum dot in each of the four corners where two elec- trons populated in antipodal sites due to the Coulomb repulsion. The proposed nano-processor is only 4-bit processor and can just handle basic tasks.
Nanosensor and Nanoactuator
Graphene and its derivatives have the potential to develop many types of sensors and actuators which can measure and analysis of different, unforeseen essential parameters and magnitudes right bottom at molecules level. These sensors can be used to measure mass, pressure, vibration, the concentration of a given gas or to detect lung cancer and asthma attacks [1]. Different types of chemical and biological nanoactuators have been reported in the literature based on the inter- action between nanomaterials, electromagnetic (EM) fields and heat [1]. In this direction, a nanoactuator which is based on the magnetic nanoparticles can kill cancer cells by heating them [35].
Nano-transceiver
A sensor node requires a communication system to communicate the sensed events to a distinct receiver. Conventional antenna, of few centimeters, usually radiates at the Gegahertz (GHz) frequencies. However, scaling down the conventional metallic antenna to nanoscale requires high operating frequency (more than 100 THz) [36]. This approach is not feasible because, at so high operating frequencies, these nodes are expected to experience extremely high path loss and absorption which are not practical to operate for such resource-restricted nodes. On the other hand, designing nanoantennas using nanomaterials such as graphene, re- duces the operating frequencies at few THz [6, 13]. Graphene has been proposed Chapter 2 Background and the State of the Art 16 as a building material for plasmonic nano-transceivers [13]. Graphene can sup- port propagation of tightly confined Surface Plasmon Polaritons (SPPs) in the ter- ahertz band (0.1-10 THz) at room temperature [37], enabling the miniaturization of nanoantenna suited for wireless communication among nanoscale nodes. The nano-transceiver will generate surface plasmonic polariton (SPP) wave which leads to femtoseconds long pulse [12, 13, 18, 19]. The nanoantenna converts the SPP wave into EM waves and radiates it into the free-space [13, 18, 19].
Nano-memory and Nano-capacitor
Nano-memory is the data storage unit and plays a vital role in an electronic de- vice because the device operations rely directly on the stored configuration pa- rameters on the available memory. Several types of nano-memories based on different technologies with a length of a few hundreds of nanometers have been reported [33, 38]. A nano-capacitor, on the other hand, is the energy storage unit intermittently powering the different units of a sensor node. The nano-capacitor is connected to the electrodes of the energy harvester to get charged. The maxi- mum energy stored in the capacitor depends on the capacitance of the capacitor, the area of the plates and voltage source through the well-known expression [33]:
1 E = CV 2 (2.1) max 2 c where C is the capacitance, and Vc is the voltage source. For electrostatic ultra- nano-capacitors with capacitance of 9 nF and voltage value, Vc, up to 0.4 V, the maximum energy can be store is approximately 800 pJ [12].
Energy harvesting unit
Due to the severely constrained area, sensor nodes are extremely power restricted. In recent years, researchers have developed different techniques to harvest energy from the environment using specialized nanomaterials [39]. These techniques are Chapter 2 Background and the State of the Art 17 piezoelectric [20, 21, 40], thermoelectric [28], triboelectric [22–24] and pyroelec- tric [25, 41]. in what follows, we provide detail literature of these techniques:
• Piezoelectric: Piezoelectricity, also called the piezoelectric effect, is the ap- pearance of an electrical potential across the sides of certain materials when they are subjected to mechanical stress. At the nanoscale, it is called nano- piezoelectricity. The discovery of piezoelectric nanomaterials provides a new opportunity to develop nanoscale energy harvesters called piezoelec- tric nanogenerators [20]. The nature of piezoelectricity comes from the non- centrosymmetricity in the crystal [39]. There are 32 crystal classes, and 20 of them exhibit the piezoelectric effect. These materials include Lead zir-
conate titanate (PZT ), barium titanate (BaT iO3), zinc oxide (ZnO), gallium nitride (GaN), zinc sulfide (ZnS) and many more [39]. The authors in [21] show that using piezoelectric zinc oxide nanowire arrays, nanoscale me- chanical energy can be converted into electricity. In particular, they find that a single zinc oxide of diameter 20 nm with the length of 200 nm can produce power up to 0.5 pW at one cycle of resonance [21]. Similarly, ver- tically align zinc oxide nanowire can generate 1.1 pW/µm3 [42].
• Thermoelectric: In this method, the temperature gradient is converted into electricity using Seebeck’s effect. The Seebeck effect, named after the Baltic German physicist Thomas Johann Seebeck, is the conversion of heat flow into power at the junction of two dissimilar electrical conductors [28]. This type of energy harvesting is feasible for portable and pervasive computing devices, and in environments where thermal gradients exist [29]. In par- ticular, human body heat can be converted into electricity using the tem- perature gradient between the body temperature and the external medium. However, low gradient and limited heat transfer can affect the output power efficiency [28]. Chapter 2 Background and the State of the Art 18
• Triboelectric: The triboelectric also known as the triboelectric effect is a type of contact electrification in which specific materials become electri- cally charged, due to electrostatic induction, after they come into frictional contact with a different material. Rubbing glass with fur, or a plastic comb through the hair can produce triboelectricity. Any materials which exhibit the triboelectrification effect, from metal, to polymer, and to silk can be candidates for fabricating Triboelectric Nanogenerator (TENG). However, the ability of material for gaining/losing electron depends on its polarity. For instance, the organic and inorganic films that exhibit opposite tribo- polarity are used to generate the triboelectricity [22]. TENGs can be used to harvest vibration energy [23], and to convert magnetic force variation to electricity [24]. It is further observed in [43] that using TENG, a power den- sity of 2.04 mW/cm3 equivalent to 0.002 pW/µm3 can be achieved by using micro/nano dual-scale polydimethylsiloxane.
• Pyroelectric: A pyroelectric nanogenerator is an energy harvesting device which converts the time-dependent temperature fluctuation into electric- ity by using nano-structured pyroelectric materials. Unlike thermoelectric, pyroelectric materials do not need a spatial gradient, but it requires tempo- ral temperature changes [28]. Likewise, a pyroelectric nanogenerator made from a single nanowire of zirconate titanate can be used as a temperature
sensor for detecting the change in temperature [25]. Using BaT iO3 film of 200 nm thick in pyroelectric nanogenerator, output power of 3 pW/µm3 can be generated [44]. However, in pyroelectric nanogenerator, the amount of harvested power is proportional to the rate of temperature change, which makes it directly useful for nanoscale systems [29]. For example, pyro- electric nanogenerator can be used to power nanosensor nodes deployed in catalyst sites within chemical reactors [11]. Chapter 2 Background and the State of the Art 19
So far we discussed different components of a nanosensor node. However, at the nanoscale, the amount of power harvested is many orders of magnitude lower than the power consumption of nanosensors [9, 11]. Table 2.1 shows that the power consumption of the nanosensors is many orders of magnitude higher than what is possibly be harvested by different nanoscale energy harvesters. This means that it is not possible to use the harvested energy to drive directly a nanoscale device that includes sensors, processors and wireless communication. Recent work [11, 27] advocates that a node should be composed of as few components as possible. This dissertation, therefore, motivates the need for sensorless event monitoring for IoNT systems which we will discuss in Section 2.4.
Nano sensors Power consumption Energy harvest- Harvested ing options power (pW /um3) Piezoelectric 0.5 [21], 1.1 Hydrogen sensor 1nW [45], 0.1 uW [46] [42] Pyroelectric 0.1 [50], 3 Pressure sensor 1 nW [47], 1 uW [48] [44] Triboelectric 2.1 [43] Temperature sensor 1 nW [49] Megnetoelectric 4.5 [51]
Table 2.1: Power consumption of different types of nanosensors versus power harvested at nanoscale.
2.2.2 Propagation Channel
The propagation of sensed information from a sensor node to a receiver, within the Internet of Nano-Things paradigm, is affected by the channel chemical com- positions. Recent studies confirm that these nodes may be able to communicate in the terahertz band using graphene as a transmission antenna [6]. The prop- agation model for THz band communication among sensor nodes is introduced in [52,53] which is based on the radiative transfer theory. In this section, we give details of the THz propagation model. Chapter 2 Background and the State of the Art 20
As terahertz band is the resonance frequency of molecules, communication in this band is severely affected by attenuation and molecular absorption noise. Radio communication is influenced by the chemical compositions of the medium in two different ways in the terahertz band. First, the radio signal is attenuated because molecules in the channel absorb energy in certain frequency bands. Sec- ond, this absorbed energy is re-radiated by the molecules which creates noise in the channel called molecular absorption noise. The attenuation in the terahertz band comes from the two factors which are spreading loss and molecular absorption loss. The spreading loss, which is the function of distance, comes from the signal propagation through channel whereas the molecular absorption loss comes from the molecular absorption of energy as molecules in the channel absorb energy in certain frequency bands. We assume that the radio channel is a medium consisting of X chemical species [x1,x2,.....xN ].
The effect of each chemical species xi on the radio signal is characterized by its molecular absorption coefficient Kx(f ) of species xi at frequency f . The molecu- lar absorption coefficients of many chemical species are available from the HIgh resolution TRANsmission molecular absorption database (HITRAN)[54]. Each ∈ type of molecule xi (xi X) has mole fraction mx in the medium. The medium absorption coefficient K(f ) at frequency f is a weighted sum of the molecular absorption coefficients in the medium:
Xn K(f ) = mxKx(f ) (2.2) x=1 where Kx(f ) is the absorption coefficient of individual molecule species at the fre- quency f . Figure. 2.2 shows the absorption coefficient of standard air with mean latitude in summer over the Terahertz band (0.1-10 THz) which is equivalent to 1 wave number from 3.3 to 330 cm− . The attenuation at frequency f and a distance d from the radio source is given by [11, 53]: Chapter 2 Background and the State of the Art 21
Figure 2.2: Interface of the HITRAN on the web tool. Molecular absorption coefficient can be calculated directly from HITRAN database.
4πf d 2 A(f ,d) = ∗ eK(f )d (2.3) c where c is the speed of light. Figure. 2.3 shows path loss in dB at different dis- tances in THz band At the nanoscale, the thermal noise is very low [1]. This is because the scat- tering of electrons in nanomaterial creates very low thermal noise. Thus the only noise which affects the communication between sensor nodes is the molec- ular absorption noise of the channel. The molecular absorption noise, Nabs(f ,d), which is due to the re-radiation of absorbed energy in a random direction by the molecules in the channel, is given [11, 12, 52, 55]:
− K(f )d Nabs(f ,d) = KBT0(1 e− ) (2.4) where T0 is the reference temperature 296K, and KB is the Boltzmann constant. Figure. 2.4 shows molecular Absorption Noise Power Spectral Density (PSD) at different distances in the THz band. Chapter 2 Background and the State of the Art 22
Figure 2.3: Total path-loss in dB of THz band at different distances.
Let U(f ) be the power spectral density of the transmitted radio signal at fre- quency f . The signal-to-noise (Signal-to-Noise-Ratio (SNR)) at frequency f and distance d is computed as follows [30].
U(f ) SNR(f ,d) = (2.5) A(f ,d)Nabs(f ,d)
2.2.3 Receiver
For any communication system, a receiver is a key component which receives sensed measurements and decodes the information sent by a sensor node. In this Section, we give details of the state of the art receiver designs which are energy based detector and CTMA (Continuous Time Moving Average) based detector.
Energy-based Detector (ED)
Energy-based Detector (ED) is a simpler circuit, low cost, but it has low sensitivity because it also captures the noise (see Figure 2.5). This is because it integrates Chapter 2 Background and the State of the Art 23
Figure 2.4: Molecular Absorption Noise PSD in dB/Hz at different distances in THz band. the signal energy in a single interval of time which is usually greater than the pulse duration . This significantly lowers the performance of the receiver since the useful signal power is also averaged with respect to the extra noise power as shown in Figure 2.6.
Band Pass 2 Classifier Filter ( )
Figure 2.5: Block diagram of energy based detection where the pulse energy is detected by integrating the received energy over an Integration window (Tint). Chapter 2 Background and the State of the Art 24
T int
Figure 2.6: In Energy detection, the integration window (Tint) is greater than the pulse width (duration) so the additional noise is also integrated.
CTMA based Detector
The schematic diagram of the CTMA-based detector is shown in Figure 2.7 where the received signal is passed through the bandpass filter to filter the noise out- side the THz band. The filtered signal is then squared and then passed through continuous-time integrator which is usually approximated using second order low pass filter [56]. The output of the low pass filter is treated as a continuous time function and input to the peak detector. The peak detector finds the max- imum signal energy and treated is an observation for decoding the information sent by the sensor node.
Band Pass 2 CTMA PD Classifier Filter ( )
Figure 2.7: Block diagram of a CTMA based detection to detect either peak power (pulse amplitude) or pulse energy.
CTMA based detector is more robust, and it outperforms the energy-based detector concerning sensitivity [31, 56]. In this dissertation, we, therefore, use the CTMA based detector to detect either peak power (pulse amplitude) or pulse Chapter 2 Background and the State of the Art 25 energy to distinguish the pulses being sent by the monitoring nodes.
2.3 Applications
Sensor nodes are typically used to carry out reasonable functions. These nodes sense and monitor cell level activities and events. Recently, the area of nanoscale communication has attracted the research community to design new algorithm and frameworks for IoNT applications tailored to the peculiarities of the THz band. Such developments have shown drastic effects in monitoring molecules level events [2]. In what follows, we provide some envisaged applications of IoNTs as illustrated in Figure 2.8.
2.3.1 Medical Applications
In the health care system, a sensor node can communicate with a micro-scale device. The sensor node processes the data locally and then can transmit it to the user device for analysis [57]. In this direction, there are significant rele- vant developments in recent years. For example, a nanorobot that can be in- troduced into the human body to detect the tumor without causing injury to the patients [58]. They propose an acoustic communication paradigm, NanoBee, in which the nanorobots communicate through acoustic signals [58]. Similarly, researchers at Imperial College of London create robotic pills which have the ca- pabilities to deliver drugs inside complex human body parts, i.e., small intestine where it is difficult for doctors to reach and treat it. The body of such a capsule is designed to have a tiny camera, a wireless chip, and a remote controller [59]. However, there are no specifications given about how to control and position the tiny robotic pills, but remote communication can be a potential solution. Once a nanorobot is deployed inside the human body then they need to prevent itself Chapter 2 Background and the State of the Art 26
(a)
(b)
(c)
Figure 2.8: Network architecture for the IoNT: (a) human lungs monitoring [2] (b) monitoring chemical reactions [3] and (c) plant monitoring [4]. Chapter 2 Background and the State of the Art 27 from the immune system through a chemically inert diamond exterior [58]. Fur- ther, in [60] the authors, inspired by biological systems, propose a nanoparticle system that can target the tumor using molecular pathways. They investigate that such nanoparticle system, with the potential of biological signaling and re- ceiving modules, can communicate information through molecular channels and can target 40 times higher doses of chemotherapeutics to tumors. These nanopar- ticles can be attached to the cell walls by using some bioengineering techniques such as atomic force microscopy or some form of artificial bacteria [60]. In [61] the authors investigate THz communication in human tissues and mea- sure the absorption path loss of skin tissues. They further propose the propaga- tion model for THz communication in vivo and derive the channel capacity and transmission range for different communication schemes. In the health-care sys- tem within IoNT paradigm, the Body Area Network (BAN) can be connected to in-body network [62] where nanomachines patrol in the body to take measure- ments and send it to the user device. They further investigate application re- quirements (functional and non-functional) for in-body networks. This propose is somehow different from the BAN that measure all kinds of body parameters from outside using wearable devices. In [63], the authors propose a Bioresorbable Electronic Stent (BES) that is integrated with therapeutic nanoparticles. For data communication, in-vivo and ex-vivo experiments are conducted between stent antenna (5 mm in size) and transmitting antenna (900 MHz and 20 mm in size). Interestingly enough, they show that 10 mW power can be transferred to the stent antenna when the incident power of the transmitting antenna is 1 W. Though the authors experimentally indicate in-vivo communication, however, it can be fur- ther extended if molecule level measurements are reported to the external world. This way, it can help to detect diseases at the early stage. These sensors, deployed within a health-care system, can be powered using an onboard energy harvester. Assuming a blood flow scenario, an estimated power of 1.28 pW can be generated when ZnO nanowires bend by a flow movement [33]. Chapter 2 Background and the State of the Art 28
2.3.2 Industrial Applications
Nanoscale communication can transfer a wide range of industrial applications by providing new solutions, easing the manufacturing process and enabling quality control procedures. For instance, nanosensors can be used to develop the touch ultrahigh sensitivity surfaces and haptic interface [1]. They can also be embedded in advanced fabrics to improve the air flow in the structures. The second most promising industrial application of IoNT paradigm is real-time monitoring and controlling chemical processes at the molecular level [2,10]. The authors propose network architecture for chemical reactors and analyse the reliability of terahertz band nanoscale communication [64]. Similarly, the challenges of attenuation and molecular absorption noise, within the terahertz band, are investigated in [30] by introducing the concept of frequency switching based on the predicted composi- tion of the medium over time [30]. In [10], the authors propose a novel concept to monitor chemical reactions and also improve the product selectivity of Fischer Tropsch (FT) catalysis. One of the major issues of nanoscale devices is the limited harvested power. Keeping this constraint in mind, a new self-powered sensing and communication architecture is proposed in [9]. The proposed architecture uses pyroelectric nanogenerator fitted in sensor nodes at catalyst sites where it harvests power from temperature fluctuations on the catalyst surface.
2.3.3 Environmental Applications
Nanotechnology can revolutionize the agricultural and food industry with new advancements in the molecular treatment of diseases and detection [65,66]. Nanonet- works can help in detecting toxic components and bacteria at the molecular level that cannot be detected using traditional sensing technologies. It can also be used to monitor plants processes such as plant emissions, humidity and temper- ature [1, 67]. As it has been reported in the literature, that plants emit various Herbivore-induced Plant Volatiles (HIPVs) with different concentrations based Chapter 2 Background and the State of the Art 29 on the insect type [67, 68]. In this direction, a nanonetwork within the context of IoNT, deploy in agriculture fields can monitor chemical compounds that are being realized and exchange between neighbor plants. Detecting such processes can help to reveal their life cycle and change patterns. It can also improve the agricultural systems, productivity by identifying the timing of an insect attack and the type of the attacking insect [67, 68].
2.4 Sensor powering challenges
Power supply at the nanoscale is a challenging problem. As it has been discussed previously that nanoscale energy harvesters (a.k.a nanogenerators) [20–25] can not generate enough power to transmit a large number of pulses uploading event information and sensor node identification each time events are detected in the environment. How to balance the use of energy so that it never exceeds the har- vested energy in tiny IoNT nodes is a challenging problem to solve. Therefore, novel energy-neutral event monitoring solutions are required. A key feature of the energy-neutral system is that it performs event mon- itoring without using sensors, which helps reducing power consumption of the system. There are other examples of monitoring without using sensors. These ex- amples appear at both macro and nanoscale levels. At the macro level, machine status is monitored using power consumption of machine as signature [69, 70]. Similarly, in [71], variation in the temperature is used as the signature to control the internal temperature of the induction motors. Recently an energy-neutral IoT system is proposed in [72] to monitor pollution using photovoltaic energy har- vested from the mini photovoltaic panels. Similarly, piezoelectric vibration en- ergy harvester can be used as a signature to recognize human activity by observ- ing the generated AC voltage [73, 74]. The authors in [75] proposed packet-less pulse switching paradigm for event sensing using Ultra Wide Band Impulse Ra- dio as the physical layer in sensor networks. The proposed concept addresses the Chapter 2 Background and the State of the Art 30 capacity and energy overheads of packet-based communication. Such a paradigm uses MAC-Routing frames to synchronize nodes with a receiver, however, events are localized by a receiver by observing the temporal position of a received pulse to a reference frame structure [75]. Recently, packet-less self-powered event monitoring ultrasonic pulse-based system is proposed in [76]. In the proposed method, a large number of sensor nodes are deployed on a plate-like structure which is being monitored. These sensor nodes, which form a Cellular Pulse Net- working (CPN), sense their local environment and communicate the event infor- mation to a base station via ultrasonic pulse links using harvested energy from the ambient vibration. Similar to work in [76], a Scalable Cellular Pulse Network- ing (SCPN) architecture is presented in [77] which retains the energy benefits of CPN while lowering the delivery delay for a larger area being monitored. These studies focus on IoT systems, hence, do not provide solutions to detect events in IoNT systems. Interestingly, energy harvester can also be used as a sensor at the nanoscale.
For instance, a TENG which is based on the contact electrification effect, has been used as a gas sensor [78]. Similarly, the correlation between the output voltage and temperature of nanogenerator can play the role of a temperature sensor [25]. However, these studies do not provide solutions to sense and monitor events re- motely because they do not cover the communication aspect. To power a tiny sensor node, there are two possible options. These two solutions are battery and energy harvesting. As batteries are difficult to build into and replace in such tiny sensors, researchers are developing nanoscale energy harvesters (a.k.a nano- generators) [20–25], that can scavenge tiny amounts of energy from the ambient environment to power event detection and very basic wireless communication where all data is transmitted as a series of extremely short (a few hundred fem- toseconds) pulses [19]. However, there is a mismatch between energy harvesting Chapter 2 Background and the State of the Art 31 and energy consumption of a nanonode [11, 18]. This means that it is not pos- sible to use the harvested energy to drive a nanoscale device that includes sen- sors, processors and wireless communication. To overcome such problem, recent work proposed a sensorless remote event monitoring framework known as SE- MON [9, 11]. The Sensorless Event Monitoring (SEMON) architecture consists of a node and a remote station. The node is located at the place where events oc- cur. The node consists of only two components: an energy harvester and a radio. When an event occurs, the energy emitted by the event is harvested by the energy harvester on the node. This harvested energy drives the radio to produce a pulse whose energy is equal to the amount of harvested energy. At the remote station, the receiver picks up the radio signal and uses an energy detector to measure the amount of received energy. Assuming that different events generate different amount of energy, the remote station can, therefore, use the received energy to identify various events. The limitation of the SEMON framework is that one remote station can only communicate with one node. This is due to the fact that the energy of the received pulse is determined entirely by the event and does not contain information on the identity of the node. This dissertation addresses the powering issue of IoNT sys- tem and studies two methods for energy-neutral IoNT, called eNEUTRAL IoNT that allow the sensors to transmit event information (the event type and its loca- tion) using only the amount of energy harvested from the events.
2.5 Pulse generation for tiny IoNT nodes
The key component of the event monitoring node is the pulse generator. There are other configurations which are used to generate SPPs on the metallic and graphene surface. These SPPs lead to very short pulses, just several tens of fem- toseconds long. Therefore, in what follows, we provide detail literature of config- urations that are used to generate SPPs as shown in Figure 2.9. At the end of this Chapter 2 Background and the State of the Art 32 section, we will provide a discussion to elaborate the need for designing a simple model to generate THz pulses suitable for Tiny IoNT nodes.
Kretchmann Configuration [87, 88] Attenuated Total Reflection Otto Configuration [84, 86, 87]
SPP Grating Coupling [80, 84, 85] Generation Methods Electron beam [82, 83]
Electrical Excitation [13, 81]
Optical Excitation [81]
Tip Method [79, 80]
Figure 2.9: Generating SPPs using different excitation methods.
Kretschmann-Raether Configuration
The Kretschmann-Raether method is an Attenuated Total Reflection (ATR) method, where Total Internal Reflection (TIR) takes place [88]. In this method, a metallic film is sandwiched between a dielectric, usually a prism, and the air [88]. Light is incident from the dielectric side and SPPs are excited on the metal-air interface as shown in Figure 2.10a). The parallel to interface component of incident light momenta (kx) couples to the SPP wave number (ksp) giving rise to SPP that which propagates along the interface (x direction) as it is shown by Figure 2.10a). A restriction that Kretschmann-Raether method imposes is that it can be applied only for very thin metallic films up to 70nm. Chapter 2 Background and the State of the Art 33
Otto Configuration
Otto method is similar with Kretschmann-Raether configuration but it follows different setting of metal and dielectrics, that is, an air gap is located between a metal and a dielectric (prism) [84, 86]. The incidence light is projected from the prism side and SPP wave is excited on the air-metal interface [87] as shown in Figure 2.10b). Likewise, in Kretschmann-Raether method, the parallel to inter- face component of the momentum of light must couple with the momentum of surface plasmon on the air-metal interface. Due to subwavelength character of
SPP [80], the wave number kx of incident EM wave is smaller than that of SPP wave ksp, that is, a denser medium is used to bring the coupling condition be- tween the light and SPP wave number, since the kx takes higher values in dielec- tric than in light as Eq. (2.7) states. The coupling condition for surface plasmon waves in the metallic film [80] is given by
kx = ksp, (2.6) where the parallel to interface wave number component of incident light is
√ kx = k0 εd sin(θi) (2.7) and the SPP wave number is given by [84, 85]
r εmεd ksp = k0 (2.8) εm + εd where k0 = ω/c is the wave number in free space, ω is the operation angular frequency, c is the light velocity in vacuum, θi is the incidence angle of light wave and εd, εm are the dielectric permittivity for dielectric and for metallic film, respectively. The SPP dispersion relation is given by the equation (2.8) at a Chapter 2 Background and the State of the Art 34 metal-dielectric interface [80] where the metal permittivity is frequency depen- dent (εm(ω)). The Otto method as well as the Kretschmann-Raether, work on ATR configu- ration where TIR happens on the dielectric interface. That is, the EM wave comes from a denser medium with an angle greater than critical angle [89]. The SPP resonance is sharp and sensitive to the coupling condition and incidence angle. This method is useful when direct contact with the metal surface is undesirable, for instance when sensing molecular absorption of the surface is required.
a) b)
Incident Beam Reflected Beam Incident Beam Reflected TM wave TM wave Beam