Design and performance evaluation of communication protocols in rfid systems

EFFICIENT MISSING-TAG EVENT DETECTION PRO- TOCOLS TO COPE WITH UNEXPECTED TAGS AND DETECTION ERROR IN RFID SYSTEMS .... 52 Psj Probability that the j-th tag is successfully detected 28 m

Hanoi−2023

MINISTRY OF EDUCATION AND TRAINING

HANOI UNIVERSITY OF SCIENCE AND TECHNOLOGY

HOANG TRUNG TUYEN

DESIGN AND PERFORMANCE EVALUATION

OF COMMUNICATION PROTOCOLS IN RFID SYSTEMS

DOCTORAL DISSERTATION OF TELECOMMUNICATION ENGINEERING

Hanoi−2023

MINISTRY OF EDUCATION AND TRAINING HANOI UNIVERSITY OF SCIENCE AND TECHNOLOGY

HOANG TRUNG TUYEN

DESIGN AND PERFORMANCE EVALUATION

OF COMMUNICATION PROTOCOLS IN RFID SYSTEMS

Major: Telecommunication Engineering Code: 9520208

DOCTORAL DISSERTATION OF TELECOMMUNICATION ENGINEERING

SUPERVISORS:

1.Assoc Prof Nguyen Thanh Chuyen 2.Dr To Thi Thao

DECLARATION OF AUTHORSHIP

I, Hoang Trung Tuyen, declare that the dissertation titled "Design and per-

formance evaluation of communication protocols in RFID systems" has been entirely composed by myself I assure some points as follows:

■ This work was done wholly or mainly while in candidature for a Ph.D research degree at Hanoi University of Science and Technology

■ The work has not been submitted for any other degree or qualifications at Hanoi University of Science and Technology or any other institutions

■ Appropriate acknowledgement has been given within this dissertation where ref- erence has been made to the published work of others

■ The dissertation submitted is my own, except where work in the collaboration has been included The collaborative contributions have been clearly indicated

ACKNOWLEDGEMENT

This dissertation was written during my doctoral course at School of Electrical and Electronic Engineering (SEEE) and Communications Theory and Applications Research Group (CTARG), Hanoi University of Science and Technology (HUST) I would like to thank all member of SEEE, CTARG as well as all of my colleagues in Military Science Academy (MSA) I am so grateful for all people who always support and encourage me for completing this study

I would like to extend my heartfelt gratitude to my principal supervisor Associate Professor Nguyen Thanh Chuyen for his instructive guidance and valuable suggestions

in my academic studies He gave me much help and advice during my PhD study and the preparation of this dissertation I am deeply grateful for his help I gratefully appreciate my secondary advisor Dr To Thi Thao for her constructive suggestions

I also acknowledge Associate Professor Le Doan Hoang from the University of Aizu, Japan, for their instructive comments and discussions about my research work I am also thankful to my friends and my fellow CTARG members for their discussions and comments about my dissertation

I would like to express my heartfelt gratitude to my family, wife, and children for their unwavering support throughout my PhD journey Their encouragement, patience, and understanding have been instrumental in helping me overcome the challenges and obstacles that I have encountered along the way Their love and sacrifices have been

my driving force, and I am forever grateful for their unwavering support Thank you for being my rock and my inspiration, I could not have done this without you

Hanoi, 2023

Ph.D Student

CONTENTS

DECLARATION OF AUTHORSHIP i

ACKNOWLEDGEMENT ii

CONTENTS vi

ABBREVIATIONS vi

SYMBOLS vii

LIST OF TABLES xi

LIST OF FIGURES xiv

INTRODUCTION 1

CHAPTER 1 BACKGROUND OF STUDY 6

1.1 Research Background 6

1.1.1 Introduction to the Internet of Things (IoT) 6

1.1.2 Radio Frequency Identification (RFID) Systems 7

1.2 Problem Statement and Literature Review 16

1.2.1 Anti-collision protocols/algorithms 17

1.2.2 Missing-tag Detection/Monitoring 23

1.3 Summary 25

CHAPTER 2 PERFORMANCE ANALYSIS OF HYBRID ALOHA/CDMA RFID SYSTEMS WITH QUASI-DECORRELATING DETECTOR IN NOISY CHANNELS 26

2.1 Introduction 26

2.2 System Description and Conventional Approach 27

2.2.1 System Model 27

2.2.2 Transmission Channel Model 28

2.2.3 Conventional Decorrelating Detector 29

2.3 Performance Analysis 30

2.3.1 Quasi-decorrelating Detector (QDD) 30

2.3.2 Performance Analysis of Tag Identification Efficiency 32

2.4 Performance Evaluation and Discussions 34

2.4.1 System Efficiency 34

2.4.2 False Alarm and False Detection 38

2.5 Summary 41

CHAPTER 3 ON THE DESIGN OF NOMA-ENHANCED BACKSCAT- TER COMMUNICATION SYSTEMS

42 3.1 Introduction 42

3.1.1 Related Works and Motivation 42

3.1.2 Major Contributions and Organization 43

3.2 System Model and Conventional Approach 45

3.2.1 System Description 45

3.2.2 Conventional Approach 46

3.3 Proposed NOMA-Enhanced BackCom Systems 48

3.3.1 NOMA-Enhanced BackCom: Static Systems 48

3.3.2 NOMA-Enhanced BackCom: Dynamic Systems 51

3.4 Simulation Results and Discussions 52

3.4.1 Number of Successful Backscatter Nodes 53

3.4.2 Number of Successful Transmitted Bits 57

3.5 Summary 60

CHAPTER 4 EFFICIENT MISSING-TAG EVENT DETECTION PRO- TOCOLS TO COPE WITH UNEXPECTED TAGS AND DETECTION ERROR IN RFID SYSTEMS 61

4.1 Introduction 61

4.2 System Description 62

4.2.1 System Model 62

4.2.2 Communication Protocol: Aloha, Wireless Channel Model, and Detection Error 63 4.2.3 Conventional Approach 65

4.3 Proposed Missing-Tag Event Detection Protocols 66

4.3.1 Protocol Description 66

4.3.2 Parameter Optimization under Impacts of Unexpected Tags and Detection Error 69 4.3.3 Expected Detection timeslots 70

4.4 Numerical Results and Discussions 71

v

4.4.1 False-Alarm and True-Alarm Probabilities 74

4.4.2 Performance Comparison with Conventional Protocols 75

4.5 Summary 76

CONCLUSION AND FUTURE WORKS 78

PUBLICATIONS 80

BIBLIOGRAPHY 81

APPENDICES 92

ABBREVIATIONS

1 APRC Adaptive Power Reflection Coefficient

3 BMTD Bloom filter-based Missing-Tag Detection

10 FDMA Frequency Division Multiple Access

21 SDMA Space Division Multiple Access

23 SINR Signal-to-Interference-and Noise Ratio

i

SYMBOLS

10 D2 Expected detection time slots of mRUN2 protocol

13 E[X i ] Expected number of slots that is expectedly empty in the i-th in

pre-computed frame but observed as non-empty in the i-th executed

frame

14 ξ i Power reflection coefficient of the i-th BN

52 Ps(j) Probability that the j-th tag is successfully detected

28 m Number of tags in E missing from population

32 Nfa Number of available tags detected as missing ones

33 Nfd Number of actual missing tags detected as available ones

37 N far Number of successful BNs from far subregion

45 P DD Bit error probability using DD

46 P QDD Bit error probability using QDD

IDs

48 Pd(a|i) Probability that a tags are not collided

49 Ps(a|i) Probability that a tags are successfully detected

50 Pcdma(a|i, K) Probability that a tags are assigned with a different codes of the K

codes

54 Pri Received power at the reader from the i-th BN

01

55 Pfp Probability that slots, which expectedly include a particular missing

tag, are observed as non-empty after nf executed frame

56 Pnear(r) Probability that a node of distance r belongs to the near subregions

57 Pfar(r) Probability that a node of distance r belongs to the far subregions

58 p n Probability of a BN being in the subregion specified by RI and r

59 p f Probability of a BN being in the subregion specified by RO and r

61 p i Probability that an expectedly empty slot is observed as non-empty

in the i-th frame

69 R mn Cross-correlation coefficient of matrix R

74 S Total number of slots used to detect a missing-tag event

01

83 X l Random variable for number of slots that is expectedly empty in

the l-th in pre-computed frame but observed as non-empty in the l-th executed frame

87 γth Reader’s sensitivity threshold

LIST OF TABLES

1.1 Tag characteristics 9

2.1 Simulation parameters for RFID system 34

3.1 Simulation parameters for NOMA-aided BackCom systems 53

3.2 Backscatter node’s data structure 53

4.1 A comparison of related works on missing-tag event detection 68

4.2 Simulation parameters for missing-tag event detection protocols 72

4.3 Optimal selection of Cth in mRUN1 and mRUN2, given Pta = 0.95 and m = T = 5 75

LIST OF FIGURES

1.1 The Internet of Things (IoT) Integration [14] 6

1.2 Components of an RFID system 7

1.3 RFID tag 8

1.4 A passive RFID tag having 96 bits of memory to represent an EPC num- ber Header identifies the version of EPC itself; EPC Manager number identifies an organization; Object class refers to a unique type of product produced by an EPC manager; Serial number uniquely identifies each item within an object class 9

1.5 A magnetic coupling RFID system 11

1.6 An electromagnetic coupling RFID system 12

1.7 An illustration of Tree-based protocol 13

1.8 FSA throughput for different frame sizes 15

1.9 An illustration of FSA protocol 15

1.10 An illustration of tag collision (a) and reader collision (b) 16

1.11 An illustration of FSA-based communication protocol with CE and DE 19

1.12 CDMA detector: A matched filter bank [78] 21

1.13 Decorrelating detector 22

1.14 A RFID system using NOMA 23

2.1 CDMA-based RFID system with FSA protocol 27

2.2 Transmission channel model 29

2.3 Reader structure with decorrelating detector 30

2.4 Quasi-decorrelating detector structure 31

2.5 Flowchart of simulation process to calculate BER and system efficiency 35

2.6 BER performance of QDD and DD detectors with respect to a number of tags, given Lc = 31, SNR = 7 dB ϵ = 3 36

2.7 BER comparison between DD and QDD by varying values of SNR 36

2.8 System efficiency with respect to the number of tags, given f = 32, K = 30, Lc = 30, SNR = 7 dB 37

2.9 System efficiency with respect to the number of tags, given K = 30, f = 32, Lc = 31, SNR = 7 dB 38

2.10 System efficiency with respect to the number of codes, given K = 30, f

= 32, Lc = 31, SNR = 7 dB 39

2.11 System efficiency with respect to frame size, given N =1000, K = 30, Lc = 31 39

2.12 False alarm and false detection rate with respect to the SNR in the conventional missing-tag detection protocols with DD and QDD, given N =1000, K = 15, f = 512, L = 4, Threshold = 0.3 40

2.13 False alarm and false detection rates with respect to the threshold in the conventional missing-tag detection protocols with DD and QDD, given N =1000, K = 15, f = 512, L = 4, SNR = 0 dB 40

3.1 Illustration of (a) system model, (b) time-slot structure, and (c) NOMA- aided BackCom system with M = 2 45

3.2 The structure of backscatter node with variable power reflection coefficients 51 3.3 The flowchart of TNP scheme 54

3.4 The flowchart of APRC scheme 55

3.5 The flowchart of DSP scheme 56

3.6 The comparison of different schemes in static NOMA-enhanced Back- Com systems 57

3.7 Normalized successful BNs versus channel threshold for the static sys- tems using the TNS scheme 57

3.8 Normalized successful BNs versus channel threshold for the static sys- tems using the APRC scheme 58

3.9 Normalized successful BNs versus channel threshold for the dynamic systems using the DSP scheme 58

3.10 Performance comparison of different schemes in static NOMA-enhanced BackCom systems 59

3.11 Performance comparison of DSP and hybrid APRC/DSP schemes in dynamic NOMA-enhanced BackCom systems 59

4.1 RFID system model 63

4.2 Aloha communication protocol with unexpected tags and detection error 64

4.3 Flowchart of mRUN1 protocol 67

4.4 Flowchart of mRUN2 protocol 68

4.5 Detection error probability Pde versus the number of tags in a slot 72

4.6 Theoretical and simulation results of the number of slots with respect

to the number of missing tags 73 4.7 Theoretical and simulation results of the number of slots with respect

to the detection error probability 73 4.8 True-alarm and false-alarm probabilities with respect to the detection

error probability of mRUN1 74 4.9 True-alarm and false-alarm probabilities with respect to the detection

error probability of mRUN2 75 4.10 The numbers of slots with respect to the number of missing tags of

conventional RUN, BMTD, proposed mRUN1 and mRUN2 76 4.11 The numbers of slots with respect to the detection error probability of

conventional RUN, BMTD, proposed mRUN1 and mRUN2 77 4.12 False-alarm (FA) and True-alarm (TA) probabilities with respect to

the detection error probability of conventional RUN, BMTD, proposed

mRUN1 and mRUN2 77

INTRODUCTION

Motivation

Radio Frequency IDentification (RFID) has become increasingly prevalent in IoT applications due to advancements in automatic identification technology The tech- nology provides a convenient means of identifying and tracking a huge number of different objects and devices that make up IoT networks It has also recently garnered significant research interests, prompting extensive exploration of challenging issues in

literature, such as tag anti-collision and missing-tag detection/monitoring Tag anti-

collision focuses on resolving signal collision caused by simultaneous transmission from multiple tags [1] This collision leads to incorrect signal decoding at central processing units (readers/interrogators) and renders subsequent operations of RFID systems non - functional Meanwhile, the missing-tag detection/monitoring issue is to design reliable protocols that can accurately detect/monitor whether some tags are missing [2]

In order to cope with the tag collision, many communications/access protocols have been designed for years They are usually based on different multiple access techniques, which schedule and control the order of each tag’s transmission Among them, Code Division Multiple Access (CDMA) [3, 4] is considered as one of the most promising anti- collision solutions, especially for dense RFID systems Each tag is assigned with one

of the orthogonal pseudo-codes so that multiple tags could be successfully identified at the same time In this case, Decorrelating Detector (DD) is usually implemented at the reader for signal decoding Nevertheless, the implementation of DD might also enhance the background noise according to [5], and thus, might degrade the system performance

To overcome the disadvantage of the noise enhancement in DD, Quasi-Decorrelating Detector (QDD) [6] has been studied as one of the alternative solutions This motivates

me to study and propose using QDD as one of the most efficient candidates for the structure of readers in CDMA-based RFID systems

Communication and signal processing algorithms/technologies also play an impor- tant role in mitigating the tag collision One of the well-known recent approaches is non-orthogonal multiple access (NOMA) NOMA enables multiple tags to be served at the same time/frequency resources by using the principle of successive interference can- cellation (SIC) decoding [7] This is achieved by differently designed transmitted power levels, also known as power-domain NOMA In [8], authors provide a design guideline

for the backscatter communication (BackCom) systems (including RFID) using a hy- brid TDMA and power domain NOMA In particular, backscatter nodes (BNs) i.e., tags are categorized into different regions based on their power levels Then, NOMA

is implemented for groups of tags selected from the different regions However, there are three major drawbacks to this design framework Firstly, the BNs are chosen for NOMA grouping in a random manner, which may increase the likelihood of signal decoding failure due to adverse wireless channel issues Secondly, BNs are assigned

by constant power reflection coefficients based on their locations, which is to make a significant difference in the channel gains Using these fixed settings, nonetheless, may lead to poor performance over time-varying channel conditions Thirdly, the design framework in [8] is intended for static NOMA-aided BackCom systems only In practi- cal systems, BNs may frequently enter or leave the reader’s coverage area, necessitating dynamic schemes These limitations make it essential to develop novel schemes to en- hance the performance of conventional NOMA-aided BackCom systems Therefore, this dissertation aims to address these limitations by proposing new paring schemes both static and dynamic BackCom systems

On the other hand, the issue of missing-tag detection/monitoring in RFID systems has been extensively studied in both academia and industry, but it is still a relatively new and under-investigated problem Many existing works assume a perfect system implementation that includes only expected tags that are known to RFID readers The assumption is, clearly, not practical since there also exists unexpected tags (the ones whose IDs are not known a priori to readers) in real RFID systems In such scenarios, the unexpected tags might result in more severe radio collision and wrong observations

of the status of each timeslot In this situation, previous protocols may report false alarms on the event detection Moreover, many existing studies do not consider the issue of detection error due to effects of wireless transmission and noise, which is a common phenomenon in RFID literature [9, 10, 11] In particular, when the received signal strength at the reader during a timeslot falls below a certain sensitivity threshold due to noise or multipath fading, the signal decoding is not successful even in timeslots with one tag’s response As a result, traditional missing-tag detection protocols may frequently produce false alarms for the system administrator, requiring more time and energy consumption This inefficiency and unreliability render the protocols no longer efficient and reliable Therefore, it is crucial to mitigate the impact of unexpected tags and the detection error in missing-tag event detection protocols

Objectives

The primary objectives of this dissertation is to offer a design framework for the performance enhancement of RFID systems, by considering the critical issues of (i) tag anti-collision and (ii) missing-tag monitoring

For tag anti-collision, the objective of implementing QDD at the reader structure

in hybrid Aloha/CDMA-based RFID systems is not only to improve tag identifica- tion performance, but also to overcome the disadvantage of noise enhancement in DD Another objective in this research area is to improve the performance of conventional NOMA-aided BackCom using novel user pairing schemes for static and dynamic sys- tems One scheme selects NOMA groups based on successful decoding probability, while another adjusts BN’s power reflection coefficients depending on their channel conditions to increase decoding probability

On the other hand, for missing-tag detection, the objective is to re-design conven- tional detection protocols taking the effect of unexpected tags and detection error into account It is expected that the new protocols outperform conventional ones in terms

of time (and/or) energy consumption

Research scope and methodology

Research scope

The research scope of this dissertation focuses on the efficient designs and analysis

in terms of time/energy consumption of tag anti-collision and missing-tag detection protocols in RFID systems The designs are mainly based on assumptions/requests of practical RFID models that might not be optimal for current protocols They, then, adopt common and different communication/signal processing technologies, algorithms

to improve the protocols’ performance It is also noted that the research scopes are contributed in terms of theoretical points of view without testbed, experimental sys- tems

Contributions and structure of the dissertation

The dissertation has three main contributions as follows:

• The first contribution of dissertation is the analysis and evaluation of the perfor- mance of hybrid ALOHA/CDMA-based RFID systems using QDD as a multi-user detector This work is proved both in terms of analysis and computer simulations

to enhance the efficiency of tag identification under effects of wireless channel impairments This contribution has been published in:

Tuyen T Hoang, Hieu V Dao, Vu X Phan, and Chuyen T Nguyen, Per-

formance Analysis of Hybrid ALOHA/CDMA RFID Systems with Quasi- decorrelating Detector in Noisy Channels, REV Journal on Electronics and

Communications, Vol 9, No 1–2, January–June, 2019

• The second contribution of the dissertation is the proposal of a comprehensive design framework for both static and dynamic NOMA-enhanced BackCom sys- tems It includes, for static systems, Two-node pairing (TNP) and novel adap- tive power reflection coefficient (APRC) schemes For dynamic systems, a novel dynamic-sized pairing (DSP) and hybrid APRC/DSP schemes are introduced These schemes are proposed to improve the system performance in terms of the number of successfully decoded bits and the number of successful multiplexed BNs from different regions for NOMA grouping This contribution has been submitted

to IEEE Access

Tuyen T Hoang, Hoang D Le, Luu X Nguyen, and Chuyen T Nguyen, On

the Design of NOMA-Enhanced Backscatter Communication Systems, IEEE

Access, DOI: 10.1109/ACCESS.2023.3272892, (ISI), May 2023

• The third contribution of the dissertation is proposal of two protocols, namely mRUN1 and mRUN2, to address the issue of missing-tag event detection These protocols use tracking counters at the reader and tags to mitigate detection errors and announce the event only if the counters reach a predefined threshold The protocols are validated through performance analysis and simulations, showing their superiority over conventional protocols in terms of false-alarm and true-alarm probabilities The contribution has been published in Wireless Communications and Mobile Computing, 2019

Chuyen T Nguyen, Tuyen T Hoang, Linh T Hoang, and Vu X Phan

(2019), Efficient missing-tag event detection protocols to cope with unexpected

tags and detection error in RFID systems, Wireless Communications and Mo-

bile Computing, DOI: 10.1155/2019/6218671, (ISI), 2019

The dissertation consists of four chapters and is organized as follows:

• Introduction provides the main motivations, objectives of the dissertation as well

as research scope, methodology, contributions, and structure of dissertation

• Chapter 1 presents the research background, problem statements, and literature review

• Chapter 2 addresses the tag anti-collision issue with CDMA-based approach in which Quasi-decorrelating detector is adopted in hybrid ALOHA/ CDMA-based RFID systems

• Chapter 3 focuses the design and analysis of efficient user pairing schemes for NOMA-enhanced Backscatter communication systems

• Chapter 4 considers the missing-tag monitoring issue with unexpected tags and detection error

• Conclusion and future works summarize the contributions of this dissertation, and introduces some future works

Chapter 1 BACKGROUND OF STUDY

1.1 Research Background

1.1.1 Introduction to the Internet of Things (IoT)

The growth of the Internet of Things (IoT) has been spurred by recent advance- ments in wireless communication and smart device technologies This enables millions

of physical objects to be connected to the Internet with widespread sensing and com - puting abilities, as illustrated in Fig 1.1 The IoT plays a crucial role in the future of the internet and has been widely discussed by both academic and industrial commu- nities because of its ability to provide various customer services in various aspects of daily life [12, 13] The IoT allows for smooth communication and automatic control among different devices without human interaction This results in the potential for industry revolution and significant benefits to society through advanced, intelligent, and automated remote management systems

Figure 1.1: The Internet of Things (IoT) Integration [14]

On the other hand, the fast growth of smart devices in IoT applications poses challenges of autonomy, low-power consumption, low cost, and high scalability [13, 15], while the essential issue is making a full interoperability of interconnected IoT devices possible [16] To cope with the issue, there are several solutions in the literature

which are Internet Protocol version 6 (IPv6), Wireless Sensor Networks (WSNs), and Radio Frequency Identification (RFID) Nevertheless, it is infeasible and un-necessary for each among billions of IoT devices to be associated with an IP address [17] In addition, the WSNs face with not-easy-to-solve challenges of energy consumption and configuration [18, 19] In this case, the RFID technology has been considered as one of the most potential candidates for the aforementioned connectivity issue in IoT thanks

to a number of considerable benefits, which is also our focus in this dissertation Indeed,

it is reported in [19] that RFID has become one of the fundamental pillars that enable the IoT and promote its rapid development It has been widely used in a variety of large-scale applications, including inventory management, logistics tracking, precision agriculture, etc., to enable automatic identification and tracking of tags attached to objects [20, 21, 22, 23, 24] Moreover, RFID is capable of detecting/identifying multiple objects/tags without line-of-sight signal propagation, while it is able to offer a number

of other great advantages including low manufacture costs (5 cents per tag), easy implementation, long service lifetime, and robustness [20]

1.1.2 Radio Frequency Identification (RFID) Systems

Radio Frequency Identification (RFID) is a contactless automatic identification and data capture (AIDC) technology that uses RF signals for communication Data is stored on silicon chips (tag memory), which are attached to targets such as books, parcels, humans, animals, or other objects This section presents the components of an RFID system and explains how it operates to provide a better understanding of RFID 1.1.2.1 RFID Components

As illustrated in Fig 1.2, a typical RFID system consists of a reader, multiple tags/transponders, and the middle-ware software (application) [25]

Data Clock Energy

(coil, microwave antenna)

Figure 1.2: Components of an RFID system

RFID Tag: Tags (or transponders) are the actual data-carrying devices that are

attached to objects for identification A tag harvests energy from reader interrogation,

Contactless data carrier = Transponder

Reader

performs lightweight computation, and transmits data in response to reader queries Due to their simple structure, small size, and low manufacturing cost, tags serve as an economical and competitive method for managing massive objects, such as inventory control, object tracking, activity monitoring, authentication, localization, and more

Figure 1.3: RFID tag

The basic components of a tag include a microchip containing non-volatile memory and an antenna to collect and transmit radio waves as shown in Fig 1.3 [26] The chip contains circuitry that stores a unique binary number in called an electronic product code (EPC) [27], while the antenna serves as the receiver and transmitter of inform a- tion EPC is a universal identifier (normally, 64 or 96 bits) that provides a unique identity to a specific physical object The antenna, which is much larger than the microchip, typically consists of loops or coiled wire extending out from the chip It receives signals from an RFID reader and backscatters the signal with required data RFID tags can be broadly classified in three types: passive, active, and semi-passive[1]

• Passive tags do not have their own power source, and they rely on the energy

emitted by the RFID reader to power them up When the RFID reader sends

a signal to the tag, the tag absorbs the energy and uses it to transmit the data back to the reader Passive tags can be further classified as low-frequency (LF), high-frequency (HF), and ultra-high frequency (UHF) tags LF tags are suitable for short-range applications, such as access control systems, whereas HF tags are ideal for mid-range applications, such as payment systems UHF tags are suitable for long-range applications, such as inventory management and supply chain management Fig 1.4 shows the EPC tag data structure of the 96-bit passive tag [28, 29]

• Active tags have an onboard power source, usually a battery, and are equipped

with a powered receiver and transmitter This enables the reception of very weak

Figure 1.4: A passive RFID tag having 96 bits of memory to represent an EPC number Header identifies the version of EPC itself; EPC Manager number identifies an organization; Object class refers to a unique type of product produced by an EPC manager; Serial number uniquely identifies each item within an object class

signals and transmission of signals over long distances or through interference Moreover, active tags are capable of detecting collisions and sensing the channel, thereby improving their overall performance They are particularly useful in sce- narios where the tag needs to operate in harsh environments or over a long range Active tags are often larger and more expensive than passive tags due to their additional components and power source However, they offer greater flexibility and functionality in terms of their communication capabilities

• Semi-passive tags combine elements of both active and passive tags They pos-

sess an onboard power source that is used to power the microchip and a passive receiver The semi-passive tag communicates using backscatter and can commu- nicate over a longer range than passive tags Semi-passive tags are typically used

in applications where longer read ranges are needed, such as in logistics or as- set tracking The use of a battery allows for the tag to transmit data at higher power levels than passive tags This feature is particularly useful in applications where tags are embedded in metal or other materials that can interfere with the radio signal Semi-passive tags are also capable of sensing their environment and collecting additional data such as temperature, humidity, or motion

Table 1.1: Tag characteristics

Multi-tag collection 3 sec to identify 20 tags 1000 tag/sec at 100 mph 7 tags/sec at 3 mph

In brief, Table 1.1 technically summarizes various characteristics a according to tag types [1, 30]

An Reader or interrogator is an electronic device that communicates with RFID

tags using radio waves The reader consists of a radio frequency module, an antenna, and a processor that performs the tag detection and data communication functions The reader sends out a radio signal, and when a tag is within range, the signal is reflected back to the reader’s antenna The reader then processes the data received from the tag and sends it to the middle-ware software for further processing [26] RFID readers can be either fixed or mobile, depending on the application require- ments

• A fixed RFID reader is a stationary device that is typically mounted in a fixed

location, such as a wall, desk, or portal It is designed to read RFID tags as they pass by the reader’s antennas These readers are designed suitable for indoor ap- plications with a low-to-moderate traffic of tagged items Fixed RFID readers can operate on a variety of frequencies, including low, high, and ultra-high frequency, and support various communication protocols Fixed readers can be connected

to a computer or network via serial, USB, Ethernet, or Wi-Fi connections They are commonly used in applications such as inventory management, asset tracking, and access control

• A mobile RFID reader, also known as a handheld RFID reader or portable RFID

reader, is a battery-powered device The device can be carried by a person and moved around as needed, making it suitable for use in a wide range of applications Mobile RFID readers are commonly used in inventory management, asset tracking, and supply chain management applications These devices typically feature an integrated antenna, a display, and a keypad or touchscreen interface They may also include wireless connectivity options such as Bluetooth or Wi-Fi for data transmission The mobility of these devices enables real-time data collection and tracking, improving the efficiency and accuracy of business operations

Middle-ware software: is an essential component of an RFID system that enables

efficient and effective management of data collected by RFID readers It acts as a bridge between the reader and backend systems, providing a layer of abstraction that simplifies data integration and processing [31] This software also provides data management capabilities, such as storing, filtering, sorting, and routing data to different applications and systems

1.1.2.2 Operating Frequencies

There are four major frequency ranges as briefly discussed below [1, 32, 33]

• Low frequency (LF): RFID systems using low frequency tags operate at a fre-

quency range below 135 kHz and typically have a read range of up to 10 centime- ters These passive tags draw power from the reader and have a data transfer rate

of less than 10 kbps They are commonly used in animal tagging, access control, and waste management, among other applications

• High Frequency (HF): High frequency RFID technology uses 13.56 MHz frequency

and has a read range of 10 to 20 centimeters The data transfer rate is less than

100 kbps, and the tags are mostly passive These tags are suitable for applications that require moderate range and are used in various settings such as access control, item tagging, and baggage control

• Ultra-High Frequency (UHF): Ultra-high frequency (UHF) RFID tags operate in

a frequency range of 860 MHz-960 MHz and have an average read range of 5 to 6 meters, but modern larger tags can reach up to 30+ meters under ideal conditions The data transfer rate is 100 kbps, and UHF RFID systems support all three types

of tags UHF RFID technology is used in various applications, including baggage handling, toll collection, and pharmaceutical serialization

• Microwave: RFID systems that use microwave tags, also known as super-high

frequency tags, operate at 2.45 GHz or 5.8 GHz frequency range These tags have

a large read range of up to 100 meters and a data transfer rate of less than 200 kbps However, microwave systems are more expensive than other types of RFID systems They are commonly used in electronic toll collection, real-time tracking

of valuable goods, and production line tracking

1.1.2.3 Communication Principle

In order to establish the communication between tags and readers, RFID technology uses either magnetic or electromagnetic coupling, which are briefly described in the following

Reader Tag

In magnetic coupling systems, a reader generates a time-varying magnetic field by passing a high-frequency current through its coil, as shown in Fig 1.5 The magnetic field produced by the reader’s coil then induces an alternating current in the coil of the tag This alternating current is rectified to a direct current to power the tag’s microchip, which then modulates the magnetic field and sends data back to the reader [25, 1] The communication between the reader and the tag is based on the strength of the magnetic field, which decreases with distance The tag must be close enough to the reader’s coil to receive enough energy to operate, but not so close that the magnetic field becomes distorted and communication is affected [25, 34]

RF signal

Backscatter signal

Figure 1.6: An electromagnetic coupling RFID system

The electromagnetic coupling systems, on the other hand, are called backscatter sys- tems As depicted in Fig 1.6, a reader emits a radio signal whose energy is transferred

to a tag’s antenna through an electromagnetic field This causes the tag’s antenna

to resonate at the same frequency as the reader’s antenna, which in turn induces an electrical current in the tag’s antenna, which powers the tag’s circuitry The tag then responds by modulating the signal and reflecting it back to the reader The reader receives the modulated signal and decodes the information stored in the tag [25, 35] The strength of the coupling between the tag and the reader depends on several fac- tors, including the frequency of the radio waves, the distance between the tag and the reader, and the size and orientation of the antennas By adjusting these parameters, RFID systems can be optimized for different applications and environments [25, 34] 1.1.2.4 Communication Protocols

As mentioned in Section 1.1.2.1, readers and tags are the two key components of an RFID system, while they are, in any practical RFID systems, usually large Commu- nication protocols are understood as the set of communication rules/methods between one or more readers and tags to perform necessary tasks, such as identification and missing-tag detection The most widely used protocol in RFID is EPC (Electronic Product Code) C1G2, which is based on a master-slave style architecture The reader acts as the master and initiates all communication, providing power for the tags to

operate The protocol is half-duplex, requiring three major steps for a reader to access

a tag [28, 36]

• First (Query): The reader sends a Query command to tags, which includes in- formation such as the session, target, and Q value The session determines the level of security and the target identifies specific tags to be accessed The Q value determines the number of (time) slots that the tag should wait before responding

to conserve energy

More generally, current communications protocols are mainly based on the principle

of time division multiple access (TDMA) in which timeslots are used to control the

number of tags’ responses They are classified into two types i.e., tree-based and (frame slotted) Aloha-based

No response

Identified

Collision

Figure 1.7: An illustration of Tree-based protocol

In tree-based protocols, collided tags i.e., the tags respond to a reader at the same timeslot, are divided (by the reader) into smaller subsets until there is only one tag

Random number

Timeslot 1 (A,B,C)

Timeslot 2 (A,C)

Timeslot 7 (B)

Timeslot 4 (A)

Timeslot 5 (C)

to increased latency and reduced system performance A simple tree structure of this kind of protocol is depicted in Fig 1.7 with three tags, i.e., A, B and C

On the other hand, in frame slotted Aloha (FSA) protocols, multiple tags can ran-

domly respond to the reader in a frame of size f of fixed-length timeslots [26, 28, 37, 38]

It is a random access approach that enables a tag to communicate with a reader in a designated slot without any predefined order In particular, at the start of each frame,

the reader sends a query command to the tags, broadcasting the parameters ⟨ f, R⟩ , where R is a random seed Upon receiving the query command, each tag selects a slot

in the frame to reply to by computing H(ID, R) mod f , where H (.) is a hash function

that has been pre-deployed by the reader and tags, and ID is the tag’s identification

This generates a slot counter SC that ranges uniformly in [1, f − 1] The reader then

processes every slot sequentially by broadcasting a “Slot end” command to terminate the current slot and begin the next one After receiving the command, each tag decre-

ments its slot counter SC by one In any given slot, when a tag’s slot counter is equal

to 0, it can transmit its messages to the reader Once a frame is executed, the reader has the ability to categorize each slot into three distinct types These include: (i) an

empty slot without any tag replies, (ii) a singleton slot with only one tag reply, and (iii) a collision slot with multiple tag replies

Here, if we assume there are totally N tags in the considered systems, the probability that k tags transmit simultaneously in one slot can be expressed as

N 1 k 1 N −k

Equation 1.1 reflects a fact that the average number of identified tags per timeslot

during the frame or the throughput (defined by µ) in other words, can be found as

plotted in Fig 1.8 for different values of the frame size

f

k

Tags’ response

Figure 1.8: FSA throughput for different frame sizes

The FSA protocols have a number of advantages over other RFID protocols For examples, they are simple, making them easy to implement and use, while they can provides high throughput and efficiency They are also our focus in this dissertation However, one of the main limitations of FSA protocols is determining the optimal frame length for a given system, especially for those with unknown and huge number of tags This requires balancing the need for high throughput with the need to avoid collisions, minimize idle/empty timeslots, and also hardware constraints [32]

E: Empty S: Singleton C: Collision

Figure 1.9: An illustration of FSA protocol

Fig 1.9 shows a simple example of Aloha-based protocol with a reading round,

where the frame size f = 5 and there are five tags Each tag randomly selects one

of five timeslots to transmit their information The resulting state of timeslots 1 and

4 are empty slot, indicating that no tags selected them Timeslot 3 is singleton slot because it is occupied by tag 3 Timeslot 2 and 5 are occupied by two tags, resulting

in a collision slot

Tag 1 Tag 2 Tag 3 Tag 4 Tag 5

1.2 Problem Statement and Literature Review

In current RFID communication protocols such as tree- and Aloha-based, signal collision usually occurs in timeslots especially when the numbers of readers and/or tags are large This causes a system’s efficiency degradation in terms of time and

energy consumption This can be classified into two types, which are tag collision and reader collision On one hand, tag collision refers to a situation where multiple

tags within a reader’s communication range attempt to transmit data to the reader simultaneously, as depicted in Fig 1.10 (a) The reader will receive a mixture of the tags’ signals and may not be able to decode the received signals correctly This leads

to re-transmissions, which can cause delays in identification of tags, waste energy and reduce the throughput/efficiency performance, especially in large-scale dense RFID systems [1, 39]

Overlapping interrogation zone

Figure 1.10: An illustration of tag collision (a) and reader collision (b)

On the other hand, reader collision occurs when two or more RFID readers located

in close proximity to each other share the same frequency channel and survey area, and send simultaneous requests to the tags located within that common area [40, 41, 42, 43]

In this case, tags cannot distinguish requests from readers, resulting in the inability

to send information back to the readers As shown in Fig 1.10 (b), tag 2 is within the common interrogation zone of Reader 1 and Reader 2, which is called overlapping interrogation zone Due to the severe performance degradation caused by the signal collision, most of current works in the literature of RFID focus on designing efficient anti-collision protocols/algorithms

Tag 2

Interrogation zone

of Reader

The deterministic anti-collision algorithms: The reader splits collided tags by send-

ing request commands using the tags’ IDs These methods are based on tree-based anti-collision algorithms such as Binary Tree (BT) [1, 26, 39] and Query Tree (QT) [1, 47] algorithms BT algorithms continue to split each subset of collided tags into two smaller subsets by a unique binary code, and query each subset until all tags have been identified QT Algorithms are similar to the BT ones, but instead of assigning binary codes to the tags, it uses a query tree structure to group tags based on certain attributes, such as their EPC prefixes Different improved version of these BT and

QT protocols have been also studied for years In particular, an advancement of BT algorithm i.e., Adaptive Binary Tree Splitting (ABTS) is proposed in [48] in which not

only collisions but also needless idle slots are decreased In [49], Chen et al introduce

a modified version of the ABTS, known as Enhanced Binary Tree Splitting (EBTS)

To reduce the receiving time at the reader, EBTS determines where the collided bits are and truncates unnecessary data bits In addition, extensive studies for the QT algorithm have been proposed in [50, 51], such as Adaptive QT (AQT) In order to

minimize collisions and expedite tag identification, the protocol utilizes information obtained from the previous reader readings In [52] another novel tag anti-collision

algorithm based on M -ary query tree scheme (MQT) is investigated This work can

achieve two objectives: firstly, eliminating unnecessary queries, and secondly, splitting collided tags into multiple smaller subsets, allowing for efficient utilization of the af- fected bits The results of this study demonstrate an out-performance of the existing QT-based algorithms

The probabilistic anti-collision algorithms: control tags’ responses in a probabilistic

(random) manner in timeslots [53, 54, 55] The FSA, which is known as the most ef- fective probabilistic algorithms and widely used in RFID standards, organizes a frame

of multiple timeslots, with each tag transmitting its ID only once per frame to mini- mize collisions This process is repeated until timeslots with the signal collision are no longer detected The principle of FSA is that if the frame size is equal to the total num - ber of tags, the number of singleton slots, and thus, the identified tags in the frame is maximum It refers to a fact that if the tag cardinality is estimated accurately, the per- formance of FSA protocol can be improved Therefore, most current protocols/works that are based on FSA try to estimate the total number of tags using observations of timeslots during frames, which are, for examples, Vogt method [56, 57], Maximum a posterior (MAP) [58], and Bayesian inference [59] In particular, two estimation meth - ods were introduced Vogt, named Vogt-I [56] and Vogt-II [57] While Vogt-I method

calculates the lower bound for the estimate as (S + 2C), where S and C represent

the observed numbers of singleton and collision slots, Vogt-II employs Chebyshev’s inequality to minimize the Euclid distance between the observed and expected vectors

of empty, singleton, and collision slots Also, in [58], the author try to maximize the posterior probability of the expected slots using their observations in each frame Although these works have proved their performance improvement, it is noted the estimation accuracy might be heavily affected by wireless fading environments with the

presence of the so-called capture effect (CE) [60, 61, 62, 63, 64, 65, 66] and detection error (DE) [9, 10, 11] or when the number of tags is significantly increased Specifically,

CE refers to a phenomenon in which a tag might be identified in a collision slot since its received SINR is higher than the reader’s sensitivity threshold In addition, in DE,

a tag might not be detected even in a singleton slot because its received SNR in this case, might be less than the threshold Figure 1.11 describes a simple example of FSA-

based protocol with the CE and DE The request f = 5 is broadcasted to all tags, and

each tag selects a slot at random to respond Here, tag 5 is supposedly detected in the

E: Empty S: Singleton C: Collision

Figure 1.11: An illustration of FSA-based communication protocol with CE and DE

collision slot 1, while tag 4 is not observed in the singleton slot 3 These phenomena result in erroneously observed slot states In other words, the timeslot observations may not accurately reflect the actual number of responses Tags can be hidden during the identification process by either another tag or an unsuccessful detection [67] As a result, the estimation accuracy of conventional methods may be degraded Extensive research on these phenomena has been conducted in the RFID literature, covering both theoretical [60, 63, 68] and experimental aspects [67, 69, 70] To cope with the cardinality estimation in the presence of the CE, several works have been proposed [9, 65, 66, 71, 72] In [65], by assuming the existence of the CE, authors try to find

a capture probability and the tag cardinality by minimizing the norm-2 distance as

in Volt-II method, which is known as capture-aware backlog estimation (CMEBE) method This work is then improved by [9] and [71] with maximum likelihood (ML) and Bayesian approach, respectively Furthermore, the work in [66] addresses the issue

of cardinality estimation in the presence of both the CE and DE using the Expectation - Maximization approach

Besides, in dense RFID systems with a huge number of tags, the estimation usually inaccurate due to hardware constraints of the frame size To cope with this problem, another anti-collision algorithm, called Probabilistic Dynamic Framed Slotted Aloha (PDFSA) has been proposed in [73] PDFSA divides the total number of tags into sub-groups by a power control approach, while DFSA is applied for each sub-group during reading rounds

1.2.1.2 Reader scheduling

Reader scheduling refers to designed processes of scheduling and coordinating the activities of multiple readers to minimize the reader collisions The objective is to

Tag 1 Tag 2 Tag 3 Tag 4 Tag 5

frame

Tag 1 Tag 5

Σ

reduce the number of required frequencies while minimizing the time/energy needed for all readers to communicate with their respective tags within their interrogation zones

1.2.1.3 Communication and Signal processing algorithms/technologies

Communication and Signal processing algorithms plays an important role in mit- igating the signal collision issue in large-scale/dense RFID systems Different works have been studied, which are based on different approaches such as CDMA-based, non-orthogonal multiple-access (NOMA), Compressed Sensing (CS), Multiple Input Multiple Output (MIMO) RFID, and hybrid systems

CDMA-based approach is well known as one of the best solutions to cope with the signal collision in conventional multiple access wireless networks Thus, it can be utilized in designing RFID tag anti-collision protocols/algorithms [45, 74, 75, 76, 77]

In these designs, each CDMA tag is assigned a distinct signature waveform, also known

as a pseudo-random (PN) code (usually, Gold code) in which each waveform c q (t) is

represented as follows

L c

n=1 where a q (n) is a PN code sequence consisting of L c chips that take values (±1) p(t)

is a pulse of duration T c , and T c is the chip interval

The characteristics of PN codes are commonly defined by their auto-correlations and cross-correlations Auto-correlation is a measure of how similar a code is to time shifted versions of itself, and cross-correlation is a measure of how similar a code is to time shifted versions of other codes in a code-set [74] The cross-correlation function

ρ i,j of a PN code, with a period of T s are given by

∫ T s

When the cross-correlation ρ i,j is zero, the codes are called orthogonal

To detect signal transmitted from CDMA-based tags, an appropriate CDMA detec- tor is implemented at the reader side The CDMA detector might consist of a matched filter bank, which is depicted in Fig 1.12 [78] where each matched filter corresponds

to one signature waveform It is assumed that Q CDMA-based tags with different

waveforms/PN codes are simultaneously transmitting signals to the reader Then, the

received signal r(t) at the reader can be represented as follows

ρ i,j =

Matched Filter Bit decision

z q = A q b q + A j b j ρ jq + n q (1.6)

j̸=q

The reader, then, can uses the knowledge of the PN codes to recover the transmitted information from each tag Nevertheless, the PN codes are usually not orthogonal in practice, which results in the so-called the multiple access interference (MAI) (the second term of (1.6) Several methods have been proposed to mitigate the effects

of MAI in which Decorrelating Detector (DD) [74] is well known as one of the most efficient solutions The structure of DD is plotted in Fig 1.13, which is composed of a

filter bank that matches the tags followed by an inverse correlation matrix R−1 Then,

the signal vector after MAI elimination denoted by bˆ can be presented as

where R is a Q × Q correlation matrix A is the diagonal matrix containing corre-

sponding received amplitudes Here, it is noted that R+ = (R −1)qq > 1 in general It

refers to a fact that although DD can completely eliminate MAI, it also enhance the background noise that might degrade the system performance

Figure 1.13: Decorrelating detector

On the other hand, NOMA allows multiple tags to be served/responded at the same time/frequency resources [7] thanks to the principle of successive interference cancellation (SIC) at the receiver of the reader side The SIC technique is based on removing the interfering signal from the received signal, one at a time as they are decoded In particular, tags that respond to the reader at the same time are required

to have different transmitted power levels, which is known as power-domain NOMA The reader first decodes the strongest signal, while treating the weaker ones as the interference After the strongest signal has been decoded, its information is used to subtract the signal of the particular tag from the received signal The decoding can only be successful if the signal-to-interference-and-noise ratio (SINR) of the considered signal satisfies a predefined threshold of the reader The reader is equipped with SIC is shown via a simple model as in Fig 1.14, where it tries to decode received signal from two tags, i.e., Tagm and Tag n Here, Tagm is assumed to experience a better channel

gain than Tagn In this case, the reader first decodes the signal of Tagm, removes the signal by SIC, and then decodes the signal of Tagn

Recently, in [8], authors provide a design guideline for the BackCom systems (that include RFID) using a hybrid TDMA and power domain NOMA In [79], a mechanism

to reduce the collisions in NOMA-aided BackCom systems is introduced by designing efficient transmitted power levels for tags The authors in [80] address the performance

of NOMA-aided BackCom systems with multiple antennas

Figure 1.14: A RFID system using NOMA

Compressed sensing is a signal processing technique that enables the efficient recov- ery of sparse signals from fewer measurements In RFID systems, compressed sensing can be used to address the challenge of reading a massive number of tags with multiple readers This approach utilizes a fact that the number of tags in practical systems is sparse (much smaller) in comparison with the whole tags’ ID space Several outstand- ing contributions have been presented in [81, 82, 83]

Multiple Input Multiple Output (MIMO) involves the use of multiple antennas at the reader and the tags to transmit and receive signals By using multiple antennas, MIMO can exploit the spatial diversity of the signals, enabling the identification of multiple tags simultaneously The use of MIMO in RFID systems to cope with collision has been also an active research area in recent years [84, 85, 86] Nevertheless, the design of the readers and tags becomes more complex and expensive, which might be unpractical/challenging in large-scale/dense RFID systems

1.2.2 Missing-tag Detection/Monitoring

Another issue, which is important, practical but independent from the previous anti-collision one is missing-tag detection/monitoring It refers to designing efficient, accurate detection mechanisms of missing tags in practical systems with a certain predefined requirements such as reliability, time and/or energy consumption [87, 88,

89, 90, 91] This issue is especially important for product management/monitoring

in large scale/dense RFID-based systems where manual checking is usually impossible [88, 92]

The missing-tag detection protocols/algorithms can be classified into two categories:

exact detection and probabilistic detection The former aims to precisely identify which

tags are missing, whereas the latter seeks to detect the occurrence of missing tags with a predefined probability While exact detection protocols provide more accurate results, they also come with higher overheads compared to probabilistic detection protocols Both types of protocols have their own strengths and values, and they are complemen- tary to each other Therefore, they should be used in conjunction with each other to provide comprehensive missing-tag detection coverage

The exact detection protocols [89, 93, 94] rely on the reader’s prior knowledge of all tag IDs and usually require tags to respond with framed slotted Aloha protocols The reader can compute the expected status (i.e., empty, singleton, and collision) of each timeslot based on the known IDs and create a bitmap By comparing the actual and expected bitmaps, missing tags can be identified For instance, if the reader does not receive a response in an expected singleton slot, the corresponding tag can be considered missing One of the major limitations of the above protocols is their inefficiency in verifying whether a tag is present in a given set Specifically, each tag must transmit its ID to the reader, which is usually 96 or 128 bits in length Consequently, when the number of tags is large, the total execution time of these protocols is extremely high

On the other hand, in probabilistic detection, the event of missing tags is detected with a predetermined probability when the number of missing tags exceeds a specified tolerance threshold One of the first protocols called Trust Reader Protocol (TRP) [93] detects missing tags by comparing pre-computed slots with those selected by the tags

in an assumed FSA-based scheme To enhance the efficiency of TRP, Multiple-Seed Missing-tag Detection protocol (MSMD)[2] uses multiple seeds to generate multiple mapping from tags to timeslots This lead to increase the probability of singleton slots, which improves detection efficiency and achieves an energy-time trade-off These traditional protocols, nevertheless, assume a perfect RFID environment with a very limited practical constraints One of the important factors that significantly affect to these protocols is the existence of unknown tags whose identities are not known to RFID readers In this case, the protocols failed to prevent the interference from the unknown tags, resulting in either reducing the efficiency or sacrificing the identifica- tion reliability To deal with this issue, in [95], an RFID monitoring protocol with UNexpected tag (RUN) is proposed in which the cardinality of unexpected tags was estimated, which is used in the missing tag detection to guarantee a required detection reliability Nevertheless, when the number of the unexpected tags is large, the perfor-