A Thesis Entitled Compact Wire Antenna Array for Dedicated Short

Home , Driven element, Loop antenna, Monopole antenna, Patch antenna, Reference antenna, Whip antenna

A Thesis

entitled

Compact Wire Antenna Array for Dedicated Short-Range Communications: Vehicle to

Vehicle and Vehicle to Infrastructure Communications

Michael A. Westrick

Submitted to the Graduate Faculty as partial fulfillment of the requirements for the

Master of Science Degree in Electrical Engineering

______Vijay Devabhaktuni, Ph.D., Committee Chair

______Richard Molyet, Ph.D., Committee Member

______Mansoor Alam, Ph. D., Committee Member

______Dr. Patricia R. Komuniecki, Dean College of Graduate Studies

The University of Toledo

December 2012

This document is copyrighted material. Under copyright law, no parts of this document may be reproduced without the expressed permission of the author. An Abstract of

Compact Wire Antenna Array for Dedicated Short-Range Communications: Vehicle to Vehicle and Vehicle to Infrastructure Communications

Michael A. Westrick

Submitted to the Graduate Faculty as partial fulfillment of the requirements for the Master of Science Degree in Electrical Engineering

The University of Toledo December 2012

This thesis contributes to the advancement of Dedicated Short Range

Communications antenna research for automobile applications. This research is focused on implementing vehicle-to-vehicle and vehicle-to-infrastructure communications to advance both the safety and the quality of the driving experience of modern motorists.

This thesis achieves three separate goals. First, a brief literature review details the current state of DSRC antenna research and the drawbacks of antenna designs previously presented. Secondly, several common wire antennas are modeled and simulated in order to assess their potential in DSRC communications. Using an industry standard electromagnetic full-wave simulator, Ansys-HFSS, all of the antenna designs examined are determined to be lacking to some degree. In this way, the need for novel antenna designs in the DSRC frequency band is clearly outlined.

The third goal of this thesis is to present a novel wire antenna design with an omni directional radiation pattern for DSRC communications. The novel design is developed based on several of the fundamental wire antennas analyzed for DSRC, but it is a unique antenna that meets all of the criteria specified in the literature review section.

Furthermore, it out performs all the currently available antennas in one way or another. iii

In order to mount the antenna in an automobile, a double-sided feed board is proposed that utilizes a grounded coplanar waveguide to excite the central monopole of the wire antenna array, and a quarter-wave transformer to improve the overall impedance match of the antenna. Vias are then added to the board around the feed line in order to suppress parallel plate modes within the substrate. Lastly, the feed board is attached with screws to a metallic cavity designed to shield the array from electromagnetic interference from within the vehicle cabin. This increases the robustness of the design in terms of both the structural and signal integrity. The overall antenna system is simulated for operation in the DSRC frequency band. The antenna presented in this thesis is found to fulfill all the necessary criteria for vehicle-to-vehicle communications, and its performance exceeds that of DSRC antennas currently available.

For my wife Danielle, my parents, my siblings, my nieces, and my nephews.

Acknowledgments

I would like to express my deepest gratitude to my advisor Dr. Vijay

Devabhaktuni for his encouragement and honesty throughout my graduate studies. He has been a steady source of wisdom and guidance, and his abundant energy and enthusiasm has made him an amazing advisor and friend.

It has been a pleasure working with the faculty, staff, and my lab mates at the

University of Toledo, during my journey through the masters program. I gratefully thank

Dr. Richard Molyet and Dr. Mansoor Alam for serving on my thesis committee and for providing me with insightful guidance throughout the years in their respective posts of undergraduate and graduate advisor for the Electrical Engineering and Computer Science department. I appreciate the funding supports offered by the Electrical Engineering and

Computer Science (EECS) department, the National Science Foundation (NSF), and

Imaging Systems Technology (IST) in giving me the opportunity to pursue my thesis research.

I would like to sincerely thank Dr. Mohammad Almalkawi for his overwhelming support in this research, and his constant encouragement. Without his calming, confidant influence this thesis would not have been possible.

Table of Contents

Abstract ...... iii

Acknowledgments ...... vi

Table of Contents ...... vii

List of Tables ...... ix

List of Figures ...... x

List of Abbreviations ...... xiv

Chapter 1: Introduction ...... 1

1.1 Motivation ...... 2

1.2 Antenna Technology Review ...... 4

1.3 Thesis Objectives ...... 8

1.4 Thesis Overview ...... 9

Chapter 2: Background on Automobile Antennas ...... 11

2.1 Low Frequency Automobile Antennas ...... 11

2.2 High Frequency Automobile Antennas ...... 14

2.3 Current DSRC Antennas ...... 16

Chapter 3: Overview on Wire Antenna Design ...... 20

3.1 Fundamental Wire Antennas ...... 20

3.1.1 Dipoles and Monopoles ...... 21

3.1.2 Loop Antennas ...... 35

vii

3.2 Common Variations on Fundamental Wire Antennas ...... 38

3.2.1 Yagi-Uda Dipole Arrays and Folded Dipoles ...... 39

3.2.2 Helical Antennas ...... 47

Chapter 4: Novel DSRC Antenna Design and Simulation ...... 51

4.1 Novel Antenna Design ...... 51

4.2 Antenna Simulation ...... 55

4.3 Feed Board Design ...... 61

4.4 Cavity Design...... 69

4.5 System Assembly and Simulation ...... 72

Chapter 5: Conclusion and Future Work ...... 75

5.1 Conclusions ...... 75

5.2 Future Work ...... 76

References ...... 78

viii

List of Tables

3.1 Variables Affecting Helical Antenna Performance ...... 48

4.1 Parameters of a Monopole Yagi-Uda Antenna Array ...... 53

4.2 Optimized Parameters of a Novel Wire Antenna Array ...... 60

4.3 Feed Network Board Details ...... 68

List of Figures

3-1 Ansys – HFSS model of an ideal dipole antenna ...... 21

3-2 Reflection coefficient(S 11 ) of a 23.9 mm dipole antenna simulated in Ansys-HFSS

in decibels (dB) ...... 23

3-3 VSWR of a 23.9 mm dipole antenna simulated in Ansys-HFSS ...... 24

3-4 Radiation pattern of a 23.9 mm dipole antenna simulated in Ansys-HFSS in the

(a) X-Y plane and (b) X-Z plane (in decibels, dB) ...... 25

3-5 Reflection coefficient, S 11 , results for parametrically varying dipole length from

20 mm to 40 mm, simulated in Ansys-HFSS, in decibels (dB) ...... 28

3-6 Imaginary component of the input impedance, in ohms, for parametrically

varying dipole length from 20 mm to 40 mm, simulated in Ansys-HFSS ...... 28

3-7 Radiation pattern at 5.9 GHz in the (a) X-Y plane and (b) X-Z plane of a 20 mm

long dipole antenna (dashed) and a 40 mm long dipole antenna (solid) simulated

in Ansys-HFSS, in decibels (dB) ...... 29

3-8 Reflection coefficient, S 11 , results for parametrically varying dipole radius from

0.25 mm to 3.0 mm, simulated in Ansys-HFSS, in decibels (dB) ...... 30

3-9 Reflection coefficient, S 11 , results for parametrically varying dipole gap size from

0.1 mm to 10.0 mm, simulated in Ansys-HFSS, in decibels (dB) ...... 31

3-10 An ideal 11.95 mm monopole antenna above an ideal, infinite ground plane

modeled in Ansys-HFSS ...... 32

3-11 VSWR (a) and S 11, in decibels (dB), (b) of dipole (solid line) and monopole

(dashed line) antennas simulated in Ansys-HFSS ...... 34

3-12 Radiation pattern of an ideal monopole antenna in Ansys-HFSS in the (a) X-Y

plane and (b) X-Z plane, in decibels (dB)...... 34

3-13 Isometric view of an ideal circular loop antenna, fed by a 50-Ω coaxial cable,

modeled in Ansys-HFSS ...... 35

3-14 Reflection coefficient of an ideal, one-wavelength loop antenna fed by a 50-Ω

coaxial cable, modeled in Ansys-HFSS, in decibels (dB) ...... 37

3-15 Radiation pattern in the (a) X-Y plane and (b) X-Z plane of an ideal, one-

wavelength loop antenna fed by a 50-Ω coaxial cable , modeled in Ansys-HFSS,

in decibels (dB) ...... 38

3-16 Ansys-HFSS model of an ideal Yagi-Uda antenna array ...... 39

3-17 VSWR (a) and S 11, in decibels (dB), (b) of a Yagi-Uda antenna simulated in

Ansys-HFSS ...... 40, 41

3-18 Radiation pattern in the (a) X-Y plane and the (b) X-Z plane (dashed) and Y-Z

plane (solid) of an ideal Yagi-Uda six-director antenna array simulated in Ansys-

HFSS ...... 42

3-19 Ansys – HFSS model of an ideal, half-wavlength folded dipole antenna ...... 44

3-20 Reflection coefficient response of a λ/2 dipole antenna (red, solid), a 1λ folded

dipole antenna (blue, dashed) and a 1 λ loop antenna (green, dotted) simulated in

Ansys – HFSS...... 45

3-21 Radiation pattern in the (a) X-Y plane and (b) X-Z plane (dashed) and Y-Z plane

(solid) of an ideal, one-wavelength folded dipole antenna, modeled in Ansys-

HFSS ...... 46

3-22 3-dimensional view (a) and 2-dimensional view (b) of a helical wire antenna,

modeled in Ansys-HFSS ...... 47

4-1 An ideal model of a novel, compact, wire antenna array on an infinite ground

plane, designed for mobile units used in future DSRC communications ...... 53

4-2 An ideal model of a wire antenna array shown with the design variables that affect

the antenna performance ...... 55

4-3 X-Y plane gain patterns of a novel wire antenna array with a varying number of

parasitic elements, measured in dB at 5.9 GHz ...... 56

4-4 X-Z plane gain patterns of a novel wire antenna array with varying spacing

between the monopole and parasitic elements, measured in dB at 5.9 GHz ...... 59

4-5 Reflection coefficient, S 11 , of an ideal model of a novel, compact, wire antenna

array on an infinte ground plane simulated in HFSS (blue, solid) and CST (red,

dashed), in decibels (dB)...... 61

4-6 Radiation pattern in the (a) X-Y plane and (b) X-Z plane of an ideal model of

a novel, compact, wire antenna array on an infinte ground plane, in

decibels (dB) ...... 61

4-7 Model of a feed board for a monopole array with the top view (a), bottom view

(b), and transformer detailed view (c) shown ...... 63

4-8 Cross section of a grounded coplanar waveguide (GCPW) transmission line ...... 65

4-9 Transmission coefficient (red, dashed) and reflection coefficient (blue, solid)

results of the quarter wave transformer, simulated in Ansys-HFSS, in decibels

(dB) ...... 68

xii

4-10 Feed board (a) top and (b) bottom with vias added around transmission line and

screws added around perimeter ...... 71

4-11 Metallic cavity design to protect and shield GCPW feed line ...... 71

4-12 Complete, novel antenna array system with antenna array, feed board, and cavity,

shown in (a) exploded view and (b) assembled view ...... 72,73

4-13 (a) Reflection coefficient, S 11 , and (b) VSWR of the fully assembled, novel

antenna system, in decibels (dB), with respect to frequency (GHz)...... 73

4-14 Radiation pattern in the (a) X-Y plane and (b) X-Z plane of the assembled, novel

antenna array system, simulated in Ansy-HFSS, in decibels (dB) ...... 73

xiii

List of Abbreviations

CST ...... Computer Simulation Technology CW……………………Coplanar Waveguide

DAB ...... Digital Audio Broadcasting dBd ...... Decibels-Dipole dBi ...... Decibels-Isotropic DSRC ...... Dedicated Short Range Communications

EM………………… .…Electromagnetic

FCC………………… ...Federal Communications Commission

GCPW ...... Grounded Coplanar Waveguide GPS ...... Global Positioning System

HFSS………………. ....High Frequency Structure Simulator

ITS...... Intelligent Transportation Services

PCB……………… .. …. Printed Circuit Board

RKE...... Remote Keyless Entry

SDARS ...... Satellite Digital Audio Radio Service SMA ...... Subminiature Version A

TPMS ...... Tire Pressure Monitoring System

V2I ...... Vehice-to-Infrastructure Communications V2V ...... Vehicle-to-Vehicle Communications VSWR ...... Voltage Standing Wave Ratio

WAVE...... Wireless Access in Vehicle Environments WLAN...... Wireless Local Area Network

xiv

Chapter 1

Introduction

The IEEE 802.11 standard defines frequencies and communication methods for wireless local area networks (WLANs). These networks have become commonplace in homes, restaurants, and shopping malls around the world. This has led to a heavy increase in WLAN users, a larger demand for wireless technology, and a rapid expansion in the number and diversity of applications for wireless networks. In July 2010, IEEE amended the 802.11 standard with 802.11p, which regulated the frequency range of 5.85-

5.925 GHz for Wireless Access in Vehicle Environments (WAVE) and Dedicated Short

Range Communications (DSRC) [1].

WAVE and DSRC communications present two unique challenges to modern antenna designers. WAVE antennas must operate inside vehicles, which are often extremely complex environments with a multitude of different materials present.

Moreover, the significant differences between vehicle interiors, even among the variations of a single vehicle model, would require antenna evaluators to repeat the computationally intensive process of modeling an entire interior many times over. For this reason, experimental results provide a much more practical solution for WAVE antenna design when compared to computer simulation. This thesis focuses instead on

DSRC communications. Unlike WAVE antennas, DSRC antennas are typically mounted to vehicle roofs. Many vehicle roofs are similar to each other because they are made up of a metal sheet with an interior lining and several exterior coats of paint. This makes computer models of DSRC antennas much simpler and more accurate than WAVE antenna models. This thesis focuses on the design an analysis of a DSRC antenna using an industrial standard electromagnetic simulator.

In this chapter, the motivation for this research is explained in Section 1.1, and a review of antenna technology is presented in Section 1.2. The objectives this thesis aims to achieve are outlined in Section 1.3, and an overview of the remainder of this thesis is given in Section 1.4.

1.1 Motivation

DSRC band communications are composed of digital information encoded in electromagnetic waves traveling from a transmitting antenna to a receiving antenna via line of sight. Specifically, electromagnetic waves in the DSRC band must have a frequency between 5.85 – 5.925 GHz. This frequency band was allocated for DSRC by the Federal Communications Commission (FCC) in 1999 as part of planned future

Intelligent Transportation Services (ITS) [2]. The two main components of Intelligent

Transportation Services, and the motivation behind this thesis, are Vehicle-to-Vehicle and Vehicle-to-Infrastructure communications.

The primary application of Vehicle-to-Vehicle and Vehicle-to-Infrastructure communications is in safety-based warning systems. Currently, vehicles manufacturers are equipping automobiles with a wide array of both wired and wireless sensors for

2 avoiding accidents by processing data and making critical decisions with faster-than- human reaction times. Vehicle-to-Vehicle communications would allow these computers to augment their sensor networks by gathering information from sensors attached to other vehicles in close proximity. This information could be used to provide drivers with lane departure warnings, overtaking vehicle warnings, emergency vehicle warnings, and many other services summarized in [3]. Timely warnings and even automated intervention systems provided by DSRC communications would greatly increase the overall safety on modern roadways.

Vehicle-to-Vehicle and Vehicle-to-Infrastructure communications could also find their use in more convenience-based applications. By transmitting data between moving vehicles and road-side units (i.e. the ITS infrastructure), traffic patterns can quickly be gathered, organized, and analyzed by governing bodies, and this data can be used to decrease congestion on heavily traveled roads. In addition, city planners could use this tool to schedule construction projects based on necessity and evaluate the adequacy of suggested detours. Also, drivers can use this data to rapidly reroute trips to avoid unnecessary delays. Although other media for communicating traffic information currently exists, transmitting it directly to a vehicle’s onboard navigation system allows the rerouting to be done automatically via hands-free commands from the driver, which is a much safer and faster alternative to conventional media.

Although the FCC allocated bandwidth for DSRC in 1999, the technology necessary to utilize it has been slow in developing. There are two main factors that have prevented this development. First and foremost, a standard for a wireless communications protocol in the DSRC band did not exist for eleven years until IEEE

3 published standard 802.11p in 2010. Second, automotive manufacturers are inherently slow in adopting new technologies when compared to other areas of industry. This is due to the rigorous amounts of testing automobile components must undergo to prove they can withstand the extreme environments they will be exposed to. In addition, satisfying government regulations for wireless antennas adds costly development time and experimental validation expenses. Lastly, the demand and popularity of vehicle technologies was initially slow to develop and difficult to predict.

While DSRC research progressed slowly its initial introduction, in recent years it has seen a surge in interest. This is evidenced by the Intellidrive Program started in late

2010 and the large number of articles published concerning DSRC since then [4]. With the attention DSRC is now getting from both the government and academia, there is a growing need for the hardware necessary to realize the many potential benefits of

Vehicle-to-Vehicle and Vehicle-to-Infrastructure communications.

1.2 Antenna Technology Review

An antenna is a broad term for any device that accepts guided electromagnetic waves as an input and transmits these waves into free space. By definition, transmitting antennas can also be used to receive electromagnetic waves from free space as an input and output guided waves. In this manner, digital information can be encoded in electromagnetic waves and communicated between physically separated systems. The distance the information can traverse and the integrity of the signal transmitted depend on the quality and performance of the antennas utilized.

In order to determine the effectiveness of an antenna, several parameters must be evaluated. The most important factor to consider for any antenna is the frequency, or cycles per second, of the signals it will be used to transmit. All antenna properties vary with respect to frequency, but appropriate designs achieve acceptable performance for specified antenna parameters within a predetermined frequency range.

After the frequency range is established, the next parameter to consider is polarization. Electromagnetic waves are composed of separate electric and magnetic waves that are coupled together. If these two waves are in phase with each other, the combined wave will always travel along a straight line. This is referred to as a linearly polarized wave. Two special cases of linear polarization occur when the wave is oriented either perpendicularly or parallel to the reference plane. These two wave types are said to be vertically polarized and horizontally polarized, respectively. However, if the electric and magnetic components of the wave are not in phase with each other, the electromagnetic wave is said to have elliptical polarization. One special case of elliptical polarization occurs when the two components are exactly 90 degrees out of phase with each other. In this case, the wave has circular polarization.

Polarization is a vital part of antenna design because antennas can only communicate with similarly polarized antennas. For example, when an elliptically polarized wave encounters an antenna that can only receive vertically polarized waves, the antenna only receives the vertical component of the wave. Therefore, when a circularly polarized wave strikes any linearly polarized antenna, exactly 50% of the wave is actually received. Also, if a vertically oriented wave meets a horizontally polarized antenna, none of the wave is received because it has no horizontal component. Because

5 of this, transmitting and receiving antennas must have identical polarizations in order to avoid significant losses due to polarization mismatch.

Once the frequency range and polarization of the antenna is determined, the next par ameter to evaluate is the antenna’s directivity. This is a ratio of the amount of energy the antenna radiates in a given direction over the amount of energy radiated in all directions. Antennas that have high directivity are useful for long distance comm unications where the antenna’s orientation is fixed (communication satellites, television antennas) or well-controlled (weather radar). Low-directivity antennas are desirable when the antenna orientation is highly variable or the antenna must be able to receive signals from multiple antennas in many different directions (cellular telephony,

AM/FM radio, etc.).

An important parameter similar to directivity is the antenna’s gain. This is a ratio of the amount of energy an antenna radiates in a specified direction over the amount of energy that would be radiated if the power supplied to the antenna was radiated by a reference antenna. The reference antenna is typically either an ideal isotropic antenna with units of decibels-isotropic (dBi) or an ideal dipole antenna with units of decibels- dipole (dBd). These antennas will be discussed more in depth later in this chapter. In this thesis, any decibel notation that does not specify between isotropic or dipole is assumed to be decibels-isotropic.

The key difference between gain and directivity is that gain is calculated using the power consumed by the antenna where direction is calculated using all the power radiated by the antenna. The significance of this is that gain reflects inefficiencies in the antenna’s radiation capabilities that waste power when electromagnetic waves are

6 transferred from guided media to free space (i.e. heat loss, vibrations, etc.). Thus, an antenna with acceptable directionality may have a low gain in the specified direction due to an inefficient design.

Since gain varies when the receiving antenna’s angular position relative to the transmitting antenna changes, it is often helpful to graphically represent these changes to fully understand an antenna’s performance. When gain is pl otted versus angular position in such a way, the resulting graph is known as the antenna’s radiation pattern. If the magnitude of the gain does not vary for all angles, the antenna is said to be isotropic, and the result is a perfectly spherical radiation pattern. If the magnitude is constant with respect to two angular dimensions but varies with respect to the third, the antenna is said to be omnidirectional and it produces a toroidal radiation pattern. If the antenna is highly directional, it will produce a radiation pattern that has a maximum at one specific point, and the gain will decrease rapidly when moving away from this point. Omni directional and directional antennas have many uses in real-world applications, while perfectly isotropic antennas represent an ideal condition not known to exist.

The final parameter commonly used to evaluate antenna performance is the characteristic impedance, commonly denoted by the symbol Zo. Characteristic impedance is the ratio of the voltage of a travelling electromagnetic wave to its current.

When a wave travelling along a transmission line reaches a change in characteristic impedance, only a portion of the energy in the wave continues travelling along the line.

The rest of the energy is reflected back to the original source of the wave. For antennas, this means the characteristic impedance of the antenna must match the characteristic impedance of the transmission line as closely as possible to avoid energy losses in the

7 form of reflected waves. The two variables most often assessed to determine proper impedance matching are the reflection coefficient, S 11 , and the voltage standing wave ratio (VSWR). The reflection coefficient is the amplitude of the wave supplied to a port over the amplitude of the wave reflected back to the port from the system. Effective antennas have low reflection coefficients because they transmit supplied electromagnetic waves into free space rather than reflecting them back toward the source.

Unless an antenna is perfectly ideal, there is always some amount of energy reflected back to the source. This reflected wave, along with the incident electromagnetic wave from the source, can sometimes form a standing wave pattern. A standing wave, unlike a travelling wave, has specific points where the amplitude of the wave is always zero (nodes) and points where the amplitude varies between the waves minimum and maximum value (antinodes). The voltage standing wave ratio is the ratio of the maximum voltage of this standing wave to the maximum voltage of the incident wave.

Like reflection coefficients, effective antennas have low VSWR values.

1.3 Thesis Objectives

The main objective of this thesis is to present a novel antenna design that satisfies the performance parameters discussed in Section 1.2 for operation in the DSRC frequency band. This antenna design provides improved performance over the current

DSRC antennas while maintaining a low profile and a low manufacturing cost. These attributes are necessary for the antenna to be feasible for automobile manufacturers.

In addition to the novel antenna design, this thesis also presents a unique feed board design for mounting the antenna in a modern vehicle. This feed board is necessary

8 to provide structural integrity for the antenna, match the impedance of the antenna to that of the transmission line, and protect the antenna from interference from intravehicular wireless systems. Like the novel antenna design, the feed board also maintains low cost features with a low profile.

1.4 Thesis Overview

The organization of the rest of this thesis is as follows. In modern automobiles, a wide variety of different types of antennas are employed in a growing number of applications. Chapter 2 provides a brief review of these antenna types in three sections.

Section 2.1 provides a review of common low frequency (f o < 500 MHz) wireless systems in modern automobiles including AM/FM radio and Digital Audio Broadcasting among others. Section 2.2 follows this with a summary of high frequency antenna systems such as satellite navigation systems. Section 2.3 concludes the automobile antenna review with a look at current DSRC antennas, and their strengths and limitations.

The novel antenna presented in this thesis is a form of a wire antenna. Therefore,

Chapter 3 contains an overview of existing wire antenna variations and their potential for

DSRC communications. The antennas presented in this chapter are modeled and simulated for DSRC band operation. Section 3.1 begins with an evaluation of the most fundamental types of wire antennas, including dipoles, monopoles, and loop antennas.

Section 3.2 follows this with a look at several common variations on these fundamental antennas. Yagi-Uda antenna arrays are analyzed first. These are followed by folded dipole antennas. Finally, helix antennas are examined and reviewed to conclude the chapter.

Chapter 4 presents a novel wire antenna array designed for DSRC band communication. Section 4.1begins the chapter with an overview of the design process for this antenna and the variables that govern its functionality. Section 4.2 continues with the performance of this antenna. Like the other wire antennas, it has been modeled and simulated utilizing Ansys High Frequency Structure Simulator. Following the antenna design and simulation, Section 4.3 presents a novel feed board designed for mounting this antenna in the roof of a vehicle along with simulation results. After this, a metallic cavity used to shield the novel antenna array from wireless communications within the vehicle is developed in Section 4.4. Finally, the entire system is assembled and simulated in

Section 4.5, and the performance of the completed antenna is analyzed.

Chapter 5 concludes this thesis with conclusions that can be drawn from the work presented. Along with this, potential work to be done in the future as a result of this thesis is discussed.

Chapter 2

Background on Automobile Antennas

Using antennas in vehicles is not a new concept. However, implementing multiple antennas in vehicles for a wide range of applications was not common practice until very recently. For most of the 20 th century, vehicles were manufactured with only one or two antennas for the sole purpose of AM/FM radio reception. In modern vehicles, though, it is not usual to find as many as ten or more antennas simultaneous operating at their designated frequencies.

2.1 Low Frequency Automobile Antennas

Because AM/FM radio antennas have been in use for the longest period of time, there is a wealth of information regarding their performance readily available. AM radio is relegated to the frequency band between 535 – 1605 KHz, while FM radios function from 88.0 – 108.0 MHz [5]. This makes AM/FM radio communications the two lowest frequency forms of wireless communications in vehicles. Since many antenna dimensions correspond to electrical length at the desired frequency of operation, antennas

11 that operate at lower frequencies tend to be larger than those that operate at higher frequencies. Therefore, radio antennas are usually the largest antennas found on vehicles.

Since AM/FM radio antennas are the largest and most common antennas in vehicles, they are also the most easily recognized. The oldest and most common radio antennas are monopole wire antennas called whips. Wire antennas will be thoroughly discussed in subsequent sections, but since they are the oldest antennas, whips have been exhaustively researched for AM/FM applications, [6] – [8]. Recently, whip antennas have been replaced with slightly smaller antennas that car makers prefer for their appearance. Blade, or shark-fin, antennas represent the majority of radio antennas in modern vehicles [9], [10]. These are composed of vertically oriented printed circuit boards covered with a radome. Their appearance and small size allows them to be mounted to the roof of the vehicle, which limits the interference the antenna experiences from the vehicle itself.

Although blade antennas have less of a visual impact than whip antennas, they still present an unsightly protrusion from the center of the vehicle roof. Vehicle stylists wanting to avoid this feature may opt instead for antennas printed onto the windshield and/or rear windshield. These on-glass antennas have received much interest in recent years due to the immense number of potential designs that can be implemented on the large glass surfaces found in automobiles [11], [12].

While printing antennas on window glass is ideal from an aesthetics perspective, from a performance standpoint it has serious disadvantages. Most notably, antennas printed on glass are placed lower on the vehicle than other antennas, and thus, they are detrimentally affected by the vehicle to a greater degree. Also, the planar design and

12 horizontal orientation of on-glass antennas make them inherently less omni directional around the vehicle when compared to their vertically oriented counterparts. So, although on-glass antennas present a more appealing physical appearance than whips or blade antennas, their performance is notably worse.

Although AM/FM radio antennas are the largest of all vehicle antennas, they only represent one type of antenna providing one service. After AM/FM radio, the next lowest frequency antennas are used for Digital Audio Broadcasting (DAB). These antennas may be similar to AM/FM antennas, as the DAB band (100 – 400 MHz) overlaps the FM band

(88.0 – 108.0 MHz) between 100 – 108 MHz. Since DAB is currently not used in the

United States, this thesis will not provide an in-depth review of DAB technology.

Just as the DAB band overlaps AM/FM radio at the lower frequencies, it also overlaps Remote Keyless Entry (RKE) and Tire Pressure Monitoring Systems (TPMS) at the higher frequencies of the band. RKE is a wireless communication system in which a driver utilizes a transmitter attached to their key ring to communicate with a receiver on their vehicle. This can provide a number of convenient functions, such as locking and unlocking doors, opening automatic doors from a distance, and starting the vehicle.

The TPMS system is comprised of air pressure sensors on all four tires paired with wireless transmitters. These communicate with a receiver mounted near the center of the vehicle, which relays the information to an electronic control unit (ECU) mounted in the instrument panel. The ECU uses the data from the receivers to calculate the tire pressure, and it alerts the driver if an irregular tire pressure is detected. TPMS and RKE systems both operate between 300 – 400 MHz. Also, both of these systems are designed to work only at a short range in and around the vehicle. Because of this, these antennas

13 are significantly different from DSRC antennas that must operate at medium to long ranges strictly outside the vehicle.

2.2 High Frequency Automobile Antennas

The previously mentioned antennas can be grouped together and categorized as low frequency antennas for vehicular applications, as there is a significant frequency gap between the RKE and TPMS systems operating frequency (~400 MHz) and the operating frequency of the next wireless system typically found in vehicles, satellite navigation systems. Satellite navigation antennas communicate with Global Positioning System

(GPS) satellites in order to calculate the vehicle ’s location and velocity, and provide the user with directions to their destination. These systems operate with a center frequency of 1.575 GHz, and therefore, many designs, such as the one presented in [13], tend to be much smaller than the antennas discussed to this point.

When GPS antennas are discussed, they are often mentioned in conjecture with

Satellite Digital Audio Radio Service (SDARS) antennas. Although there is a sizable gap between the center frequency of GPS systems (1.575 GHz) and that of SDARS antennas

(2.3 GHz), their operation requirements are similar. Both systems must communicate with satellites as opposed to terrestrial antennas, and both are relatively new to the automotive market. The requirements for both systems are similar enough that several designs have been published that integrate the two services into a single antenna that radiates at both of the necessary frequencies [14], [15]. This is possible because these two antennas require the same radiation pattern to communicate with satellites. To elaborate on this, these antennas must be highly directional because they are receiving electromagnetic waves transmitted over an incredibly long distance (satellites in

14 geostationary orbit are located approximately 22,000 miles above the equator). This is in sharp contrast to antennas that communicate with terrestrial base stations (i.e. FM/AM radio, DAB, DSRC, etc.) as they typically require an omni directional radiation pattern and radiate over much shorter distances (less than 50 miles).

At frequencies above SDARS communications, automobile antennas are relegated to local applications, with the exception of DSRC. Bluetooth technology (2.4 GHz) is often used inside vehicles to create a link between passengers’ mobile phones and the vehicle’s onboard audio and navigation system. Some vehicles employ Wi-Fi to communicate with passengers wireless devices within the vehicle cabin, though this is less common (2.4 GHz). These antennas, like RKE and TPMS, must contend with the challenges of transmitting a short distance inside a vehicle, rather than the task of transmitting medium distances outside the vehicle that is the focus of this thesis.

Lastly, vehicle detection radar systems can be found on many high grade vehicles produced in the last few years. These systems typically use antenna arrays to detect large metallic objects. The system detects not only the presence of the object, but also the speed of the object relative to the speed of the vehicle the radar is attached to. If the relative speed of the object is within a certain range, it is assumed to be another vehicle and the driver is alerted to the second vehicle’s presence. Radar antenna arrays operate at very high frequencies (~74 GHz) and are attached to the front and rear bumpers of the host vehicle. Because of this, they have minimal interference from other wireless systems in the vehicle. Also, these antennas are directed horizontally away from the host vehicle, so they also cause only minimal interference to other systems. Unlike DSRC antennas, radar arrays are designed to operate at a very short range (~10 – 20 meters).

While the antennas discussed in this section do not encompass all antennas found in all vehicles, this list presents an accurate depiction of the antennas one can expect to find in a modern automobile. As previously stated, the DSRC frequency band (5.85 –

5.925 GHz) is much higher than the frequencies other automobile antennas operate at, with the exception of radar antenna arrays. In addition, none of the other high frequency vehicular antennas are used for medium to long distance communications with terrestrial antennas. These unique frequency and distance requirements make DSRC antenna design an interesting and active area of research.

2.3 Current DSRC Antennas

As previously stated, although the frequency band from 5.85 – 5.925 GHz has been reserved for DSRC communications since 1999, the amount of research available on

DSRC antennas is limited. The necessary characteristics for a working DSRC antenna are as follows. First, DSRC antennas need to radiate and receive vertically polarized electromagnetic waves. As previously stated, this means the electric and magnetic wave components of the full electromagnetic waves are exactly in phase with each other, and the combined wave is oriented along the vertical axis of the antenna.

A second crucial requirement for DSRC antennas is an omni directional radiation pattern in the azimuth plane. This concept is somewhat intuitive, as the antennas are designed to communicate with vehicles and roadside stations in the front, back, and along both sides of the vehicle. Therefore, there is no need to provide reception capability to the area directly above the vehicle, and the gain of the antenna should decrease as the elevation increases. Omni directionality is also important because the antenna will be

16 attached to a vehicle in motion, and it will be transmitting and receiving signals from antennas attached to other vehicles also in motion. Any nulls in the azimuth plane of the radiation pattern would result in temporary losses of communication. This is the equivalent of a blind spot for the antenna, which would severely compromise the safety features previously presented.

In recent years, a few novel DSRC antenna designs have been published. Much of the research that has been done, however, focuses on designing vehicle-to- infrastructure roadside stations rather than the mobile units. These stations can be designed using dipole and loop antennas [16], and many other types of antennas as well

[17]. While these designs achieve antennas that radiate at the desired frequency, they do not have the necessary radiation pattern for vehicular use. The purpose of DSRC roadside units is to communicate with vehicles. Because of this, these systems need to radiate only in the direction or directions that vehicles are approaching from. Therefore, these units are highly directional and oriented towards oncoming traffic. For vehicular units, this is unacceptable as previously stated.

Many of the recent vehicular designs focus on integrating the GPS antenna and the DSRC antenna, such as in [18] and [19]. While combining these antennas is an effective method for conserving vehicle space, it creates an additional concern for antenna performance. GPS antennas are typically located in the center of the vehicle under the windshield. This is an acceptable location because the satellites they are targeting are almost directly overhead. Additionally, this location is convenient because it puts the antenna in close proximity to the vehicle’s onboard na vigation system, which it must communicate with. DSRC antennas, on the other hand, are typically located on the

17 vehicle roof. This location provides a large metal ground plane (i.e. the roof) and a clear, unobstructed view around the vehicle for an omni directional radiation pattern.

Furthermore, the cost saved by combining the two antennas is negated by the added wiring for the GPS antenna, and the added complexity in the manufacturing design. This makes utilizing two separate antennas the more attractive option.

Analyzing the performance of the antennas presented in [18] and [19] reveals a few more drawbacks of the combining GPS and DSRC antennas. For the microstrip patch antenna designs, the large ground plane simulations result in a maximum gain of only 5 dBi. Also, the maximum gain for this antenna occurs at a theta value of 45 degrees, and it reaches 0 dBi around 80 degrees. For DSRC communications, the gain should be above 5 dBi in the range of 75 – 90 degrees, as the target antennas are typically located horizontally with the antennas or only slightly elevated. This low gain at 80 degrees assumes the vehicle antenna is physically much lower than the target antenna.

This works well for vehicle-to-infrastructure communications when the vehicle is near the intended base stations, but when the vehicle is further away from the base station, the angular position of the target antenna can be 80 degrees or even slightly lower.

Furthermore, this antenna would not be practical at all for vehicle-to-vehicle communications, where the target antenna is locate in the same plane as the vehicle antenna. Therefore, in DSRC communications, the gain of the antenna for low elevation angles is more important than the gain at high elevation angles.

The last concern regarding integrating the GPS and the DSRC antennas is the absence of WLAN radiation in these designs. Many antenna designs, such as [20], combine GPS and WLAN antennas to reduce costs and save vehicle space. However,

18 current antenna designs that integrate the GPS antenna with the DSRC antenna do not also integrate the WLAN antenna. As previously stated, most antenna dimensions are based on electrical length. Therefore, lower frequency antennas tend to be larger than high frequency antennas. Since GPS and WLAN frequency bands are well below the

DSRC frequency band, those two types of antennas are larger than DSRC antennas.

Because of this, it is more effective to combine the two larger antennas into one antenna and design a standalone DSRC antenna, and that is the direction chosen for this thesis.

Chapter 3

Overview on Wire Antenna Design

This chapter provides a thorough explanation of the theory behind designing wire antennas and the method used to design the novel wire antenna presented in this thesis.

Section 3.1 begins with a brief review of fundamental wire antennas and their past and present applications. The typical performance of several of these antennas will be analysed for DSRC band transmission, including dipoles, monopoles, and loop antennas.

In Section 3.2, several variations on the fundamental wire antenna types are presented and simulated for DSRC band operation. The antenna analysed in this section include

Yagi-Uda antenna arrays, folded dipole antennas, and helix antenns.

3.1 Fundamental Wire Antennas

The DSRC antenna presented in this thesis is a type of wire antenna, which is the oldest and most common type of antenna. In fact, Guglielmo Marconi achieved the first transatlantic wireless communication via the use of two enormous wire antennas [21].

There are many different types of wire antennas, including dipoles, monopoles, loop antennas, and numerous others. Modern wire antennas are much smaller and more efficient that Marconi’s, but the principles that govern them are the same.

3.1.1 Dipoles and Monopoles

One of the simplest antennas is the common dipole antenna. It is made up of two radiators oriented in the same direction, sharing a common axis, and separated by an air gap. The two radiators are excited by a feed line in the center of the antenna. A complete review of the parameters that govern the functionality of this antenna can be found in [22]. In order to evaluate dipole antennas for DSRC communications, a dipole was modeled using High Frequency Structure Simulator (HFSS) by Ansys. This model is shown in Figure 3-1.

Figure 3-1: Ansys – HFSS model of an ideal dipole antenna.

There are three main parameters that govern the operation of this antenna. In no particular order, they are dipole length, air gap, and dipole radius. The air gap and the dipole radius have only minor effects on the performance of the dipole. Because of this, their size is typically determined based on the antenna manufacturer’s requirements, within a reasonable range. That leaves the dipole length as the primary concern when designing antennas of this type.

Although the dipole has a physical length, when designing antennas it is often much more convenient to refer to the electrical length of the antenna rather than the nominal value. Electrical length is the distance from one end of the antenna to the other measured in wavelengths of the signal at a specified frequency. Thus, at differing frequencies the same electrical length refers to different nominal lengths. Electrical length is usually stated in terms of wavelengths, denoted by the greek letter lambda, λ.

The most common dipoles for practical use are half-wavelength dipole antennas.

As stated in [22], one of the primary advantages of half-wavelength dipoles is that their characteristic impedan ce is 73 Ω, which is very close to the characteristic impedance of commercially avai lable 75 Ω transmission lines. In addition, longer dipoles ( l > λ) have radiation patterns with multiple lobes, where as half-wavelength dipoles only have one lobe. A single lobe pattern is usually preferred because multiple lobe patterns can have significant nulls between the lobes, which can be extremely detrimental to antenna performance. Finally, half-wavelength dipoles provide an excellent balance by providing a more compact, material saving design than longer antennas without realizing the sharp performance reduction seen in shorter antennas. These factors have all combined to make half-wavelength dipoles an attractive and popular choice for antenna designers for many years.

In order to design the half-wavelength dipole depicted in Figure 3-1, the electrical length of λ /2 must be converted to a nominal value. This can be accomplished through equation (1).

LD = λ /2 = (0.5) c / f o (1)

In this equation, LD is the length of the dipole, c is the speed of light in a vacuum, and f o is the desired operating frequency. Substituting the middle frequency of the DSRC band

(5.8825 GHz) into the equation for f o gives a nominal ideal length of 25.42 millimeters.

However, this ideal length does not factor in the radius of the wire used. To correct for this, the length of the dipole, L D, must be shortened to between 0.47λ and 0.48λ, depending on the radius [22]. Substituting 0.47 and 0.48 for 0.5 in equation (1) gives a practical range for L D of 23.89 millimeters to 24.44 millimeters. The model used for this simulation utilized the 0.47λ value rounded to the nearest tenth of a millimeter, 23.9 millimeters.

By this point, it is quite clear that dipole antenna design is a simple process with relatively few variables. Because the length of the dipole has such a dominating effect over the other variables, the optimization process is fast and efficient, and the result is a reasonable range of lengths. This makes this type of antenna favorable for manufacturing purposes because it is unaffected by slight variations. This fact is supported by the reflection coefficient (S 11 ) shown in Figure 3-2.

0 -2 -4 -6 -8 -10 -12 -14 -16 4 4.5 5 5.5 6 6.5 7 7.5 8 Frequency (GHz)

Figure 3-2: Reflection coefficient (S 11 ) of a 23.9 mm dipole antenna simulated in Ansys- HFSS in decibels (dB).

This graph shows that this dipole antenna provides an acceptable impedance match (S 11 < -10 dB) from approximately 5.25 – 6.0 GHz. This translates to a 12.7% bandwidth, which proves that the dipole antenna has a large manufacturing tolerance.

This impedance match is further proven by the acceptable voltage stand wave ratio

(VSWR ≤ 2.0) from 5.23 – 6.0 GHz shown in Figure 3-3.

0 4 4.5 5 5.5 6 6.5 7 7.5 8 Frequency (GHz)

Figure 3-3: VSWR of a 23.9 mm dipole antenna simulated in Ansys-HFSS.

As previously stated, the wide bandwidth and design simplicity are two major advantages of dipole antennas. The third advantage is their omnidirectional radiation patterns. Due their shape and design, dipole antennas can be excited by incident electromagnetic waves from a wide range of angles. In the azimuth, a dipole is a perfectly symmetrical cylinder. This results in a completely omni directional (toroidal) radiation pattern. As the elevation angle of the incident wave changes, it alters the effective length of the dipole antenna that the wave strikes. Referring back to Figure 3-2 and 3-3, the bandwidth of the dipole antenna provides the added benefit of large gain over a wide range of elev ation angles (45ᴼ ≤ θ ≤ 135ᴼ). This is graphically represented in

Figure 3-4 below.

0° 0° 2.5 10 30° 330° 30° 330° 2 0 -10 1.5 60° 300° 60° -20 300° 1 -30 0.5 -40 90° 0 270° 90° -50 270°

120° 240° 120° 240°

150° 210° 150° 210° 180° 180°

(a) (b) Figure 3-4: Radiation pattern of a 23.9 mm dipole antenna simulated in Ansys-HFSS in the (a) X-Y plane and (b) X-Z plane (in decibels, dB).

This plot clearly shows a completely omni directional radiation pattern that decreases significantly for extreme elevation angles (θ < 45ᴼ ; θ > 135 ᴼ). In terms of

DSRC applications, the omni directional pattern is ideal, and the low gain at extreme elevation angles is not a concern as vehicle antennas are not required to communicate with anything located directly above or below them. However, the level of maximum gain this antenna achieves is insufficient. As shown in Figure 3-4, dipole antenna’s have a maximum gain of approximately 2.1 5 dB for all angles in the azimuth direction (0ᴼ ≤ φ

≤ 360ᴼ) in the X -Y plane (θ = 90ᴼ). This is much lower than the current DSRC antennas discussed in Section 2.2, and much too low to be viable for DSRC. For this reason, dipole antennas are not considered as a potential solution for future DSRC applications.

Even though it is clear that dipole antennas are not fit for vehicle-to-vehicle or vehicle-to-infrastructure communications, they are still worth analyzing in-depth. Dipole antennas are a fundamental building block for many other antennas and antenna arrays.

Therefore, it is crucial to fully understand the operation of this antenna so that antennas

25 based on it can likewise be fully understood. To this end, several parameter sweeps were performed. Namely, the dipole length, dipole radius, and the gap between the dipoles were all varied to determine the effect they have on the antenna’s performance.

The first and most dominant parameter examined in depth is the dipole length. As mentioned previously, the dipole length is the determining factor for the antennas center frequency. For the previous dipole example, the length of the dipole was determined by setting the center frequency to 5.8825 GHz and solving equation (1) to establish a nominal value equivalent to one-half of a wavelength at the center frequency. This length is chosen because it allows the ideal standing wave pattern to form on the dipole.

A standing wave occurs when waves traveling in opposite directions sum together to form a standing wave. In this type of wave, the nodes (zeroes) and antinodes

(maximums) are spatially fixed. The nodes naturally form at the end points of the dipole, because any charges that travel to the end of the antenna ar e reflected 180ᴼ back towards the source. Since the magnitude of the charges approaching the end point is always the exact opposite of the magnitude of the charges leaving the end point, their magnicvxtudes always sum to exactly zero. In a traveling sinusoidal wave, zeroes and maximums/minimums are spatially separated by exactly λ/4. Therefore, the nodes and antinodes of the standing wave formed in a dipole antenna are also separated by λ/4.

When a dipole antenna is exactly λ/2 long, the center point i s located precisely λ/4 away from both end points, and it therefore coincides with the single antinode of the standing wave. This important because it means the standing wave has the same phase for the entire length of the dipole antenna, and therefore the entire antenna contributes to the radiating electromagnetic wave. If the dipole is longer than λ/2, the antinode will not

26 be located at the feed point, and some energy will be reflected back to the source rather than being radiated by the antenna. If t he antenna is shorter than λ/2, there will not be an antinode, and again, excess energy that cannot be radiated will be reflected back to the source. Therefore, the center frequency of the antenna is the frequency where the antinode of the standing wave forms at the feeding point, and the antenna radiates with maximum efficiency. At this frequency, the nominal length of the antenna is equivalent to an electrical length of half of a wavelength (i.e. L D = λ/2).

fo = (0.5) c / LD (2)

When the length of the antenna is parametrically varied, the center frequency of the resulting antenna can be solved for by simply rearranging equation (1) to get equation

(2). As this equation indicates, the center frequency is inversely proportional to the antenna length. This observation is confirmed with the reflection coefficient results of the parametric simulation performed, shown in Figure 3-5. As noted previously, the center frequencies for these antenna lengths do not correspond to exactly λ /2, but rather they translate to approximately 0.47λ to 0.48λ. This is necessary because the ideal value of λ/2 assumes an ideal dipole radius of zero. Any real antenna has a nonzero radius, and this creates a small amount of reactance in the input impedance. In order to eliminate this reactance at the center frequency, the antenna must be shortened slightly. The amount that the antenna needs to be shortened depends on the radius and length of the dipole. In general, thicker wires will need to be shortened more than thinner wires. This can be confirmed by comparing the reflection coefficients for various dipole lengths shown in

Figure 3-5 to the imaginary components of their input impedances in Figure 3-6. The

27 center frequency of the reflection coefficient aligns with an imaginary input impedance component of zero for each dipole length simulated.

0 Dipole Length = 20mm -2 Dipole Length = 22mm -4 Dipole Length = 24mm -6 Dipole Length = 26mm Dipole Length = 28mm -8 Dipole Length = 30mm -10 Dipole Length = 32mm -12 Dipole Length = 34mm Dipole Length = 36mm -14 Dipole Length = 38mm -16 Dipole Length = 40mm -18 4 4.5 5 5.5 6 6.5 7 7.5 8 Frequency (GHz)

Figure 3-5: Reflection coefficient, S 11 , results for parametrically varying dipole length from 20 mm to 40 mm, simulated in Ansys-HFSS, in decibels (dB).

150 100 Dipole Length = 20 mm 50 Dipole Length = 22 mm Dipole Length = 24 mm 0 Dipole Length = 26 mm -50 Dipole Length = 28 mm -100 Dipole Length = 30 mm -150 Dipole Length = 32 mm Dipole Length = 34 mm

Reactance (Ohms)Reactance -200 Dipole Length = 36 mm -250 Dipole Length = 38 mm -300 Dipole Length = 40 mm -350 4 4.5 5 5.5 6 6.5 7 7.5 8 Frequency GHz

Figure 3-6: Imaginary component of the input impedance, in ohms, for parametrically varying dipole length from 20 mm to 40 mm, simulated in Ansys-HFSS.

The center frequency shift in the reflection coefficient shown in Figure 3-5 corresponds to a similar shift in the voltage standing wave ratio. The radiation pattern, however, follows a unique trend. When the dipole is tuned to the correct center frequency, as it is in Figure 3-4, the radiation pattern is omni directional as shown. If the dipole is too short, the antenna ’s gain decreases more slowly as the elevation angle increases or decreases from the azimuth plane. The resulting radiation pattern at the measured frequency appears more stretched out as shown in Figure 3-7 (a). Likewise, if the dipole is too long, the antenna ’s gain decreases more rapidly as the elevation angle increases or decreases from the azimuth plane. Because of this, the resulting radiation pattern appears fatter, as shown in Figure 3-7 (b).

0° 0° 4 20 30° 330° 30° 330° 3 0 -20 60° 2 300° 60° 300° -40 1 -60 90° 0 270° 90° -80 270°

120° 240° 120° 240°

150° 210° 150° 210° 180° 180°

(a) (b) Figure 3-7: Radiation pattern at 5.9 GHz in the (a) X-Y plane and (b) X-Z plane of a 20 mm long dipole antenna (dashed) and a 40 mm long dipole antenna (solid) simulated in Ansys-HFSS, in decibels (dB).

After the dipole length, the next parameter of interest is the dipole radius. This is because the thickness of the dipole wire determines how well the antenna radiates. As the ratio of the length of the wire to the diameter decreases, the antenna becomes more difficult to excite. This can be attributed to the additional metal and the additional

29 capacitance between the two arms of the dipole antenna. While the reflection coefficient worsens, the bandwidth of the antenna actually increases. Also, the center frequency decreases slightly with increasing radius. The worsening S11, increasing bandwidth, and decreasing center frequency can all be seen in the parametric sweep shown in Figure 3-8.

-2 Dipole Radius = 0.25 mm Dipole Radius = 0.5 mm -4 Dipole Radius = 0.75 mm Dipole Radius = 1 mm -6 Dipole Radius = 1.25 mm -8 Dipole Radius = 1.5 mm Dipole Radius = 1.75 mm -10 Dipole Radius = 2 mm Dipole Radius = 2.25 mm -12 Dipole Radius = 2.5 mm Dipole Radius = 2.75 mm -14 Dipole Radius = 3 mm

-16 4 4.5 5 5.5 6 6.5 7 7.5 8 Frequency GHz

Figure 3-8: Reflection coefficient, S 11 , results for parametrically varying dipole radius from 0.25 mm to 3.0 mm, simulated in Ansys-HFSS, in decibels (dB).

The third and final parameter affecting dipole performance is the gap between the two arms of the dipole. The air gap is necessary because it isolates the two arms from each other and forces the current to travel along the antenna and create radiating electromagnetic fields. Without an air gap, there would be a short circuit between the transmission line, and the dipole antenna would essentially be removed from the electric circuit. The effect of the size of the air gap can be seen in Figure 3-9.

-2 Dipole Gap = 0.1 mm -4 Dipole Gap = 1 mm -6 Dipole Gap = 2 mm Dipole Gap = 3 mm -8 Dipole Gap = 4 mm -10 Dipole Gap = 5 mm -12 Dipole Gap = 6 mm -14 Dipole Gap = 7 mm Dipole Gap = 8 mm -16 Dipole Gap = 9 mm -18 Dipole Gap = 10 mm -20 4 4.5 5 5.5 6 6.5 7 7.5 8

Frequency (GHz)

Figure 3-9: Reflection coefficient, S 11 , results for parametrically varying dipole gap size from 0.1 mm to 10.0 mm, simulated in Ansys-HFSS, in decibels (dB).

This graph clearly shows that the dipole gap size has a significant effect on the antenna’s center frequency. Larger air gaps decrease the center frequency while shorter air gaps increase the center frequency. This is consistent with Figure 3-5, which shows that altering the length of the dipole arms has a similar effect. In fact, both parameter sweeps are producing similar results because they are both increasing the overall electrical length of the antenna. The plot above also shows better radiation (i.e. lower

S11 ) and a decrease in bandwidth as the dipole gap increases. Along with causing a decrease in bandwidth, increasing gap size also increases the difficulty in implementing the dipole in practice. For the dipole to radiate properly, the air gap cannot contain any metallic mounting equipment, and the dipoles must be correctly aligned and completely fixed at a distance. This is much easier to achieve in practice when the gap is small than

31 when it is large. Because of this, the dipole gap is usually a fixed value, and the dipole length is used to tune the antenna to the correct operating frequency.

In summary, the three parameters that affect a dipole’s performance are the dipole length, radius, and gap size. The length and the gap size can be used to tune the antenna to the correct operating frequency, while the radius and the gap size are altered to adjust the antenna’s bandwidth. All three of them must be considered in any dipole design.

Along with the dipole antenna, another of the most popular wire antennas is the monopole antenna. A monopole can be thought of as one-half of a dipole antenna. In this design, the second half of the dipole has been replaced by an ideal ground plane.

According to electromagnetic image theory, radiated waves from the antenna reflect off the ground plane in a way that the ground plane can be thought of as a mirror image of the monopole antenna. Essentially, this means that a monopole antenna performs very similar to a dipole antenna, because the ground plane image of the monopole creates the missing half of the dipole. An Ansys-HFSS model of a monopole antenna on an infinite ground plane is shown in Figure 3-10.

Figure 3-10: An ideal 11.95 mm monopole antenna above an ideal, infinite ground plane modeled in Ansys-HFSS.

The design process explained for dipole antennas applies in the same manner to monopole antennas. However, where dipole antennas have an electrical length of λ/2, monopole antennas are exactly half that long, or λ/4. For this reason, quarter-wavelength monopoles are often used over dipoles because they are more compact. For vehicular applications, (i.e. AM-FM radio) monopoles are more easily adapted than dipoles because dipoles operate in free space. Mounting an antenna in free space is difficult, and sometimes impossible to achieve. This is especially true in mobile applications, where the mounting structure has to be exceptionally durable. Monopoles, on the other hand, need to be attached to a conducting metal plane that forms the monopole image that allows the antenna to work. In the case of vehicles, mounting a monopole antenna is not a challenge,as vehicle antennas are already typically installed on a conducting metal plane (i.e. the roof of the car). This conducting metal plane also provideds the added function of isolating the monopole antena from intravehicular antenna systems such as the tire pressure monitoring system and the remote keyless entry system.

Because of their similar designs, monopole and dipole antennas produce similar reflection coefficients and voltage standing wave ratios. Figure 3-11 shows the VSWR and S 11 from the dipole antenna previously analyzed along with those same values for monopole antenna designed to radiate at the same frequency. As these graphs show, there is little difference in the reflection coeficient (< 2dB) and VSWR within the operating band. Therefore, both antennas produce sufficient impedance matching results for DSRC communications. As before, the next step in evaluating the monopole antenna is to examine the radiation pattern, shown here in Figure 3-12.

0 -2 -4 -6 -8 -10 -12 -14 -16 -18 -20 4 4.5 5 5.5 6 6.5 7 7.5 8 Frequency (GHz) (a) 25

0 4 4.5 5 5.5 6 6.5 7 7.5 8 Frequency (GHz) (b) Figure 3-11: VSWR (a) and S 11, in decibels (dB), (b) of dipole (solid line) and monopole (dashed line) antennas simulated in Ansys-HFSS.

0° 0° 10 6 30° 330° 30° 330° 0 5 4 -10 60° 300° 60° 3 300° -20 2 -30 1 90° -40 270° 90° 0 270°

120° 240° 120° 240°

150° 210° 150° 210° 180° 180°

(a) (b) Figure 3-12: Radiation pattern of an ideal monopole antenna in Ansys-HFSS in the (a) X-Y plane and (b) X-Z plane, in decibels (dB).

This plot reveals that a monopole antenna has a radiation pattern similar to only the top half of a dipole radiation pattern. This can be explained by the ideal ground plane the antenna is mounted on. Electromagnetic waves cannot travel trough an ideal ground plane, and therefore, any waves incident on the monopole from below the horizon (i.e. θ

> 90ᴼ) will be conducted along the groun d plane instead of being absorbed by the antenna. In addition to this, any waves incident on the antenna above the horizon line will be partially reflected by the ground plane towards the antenna. The result of this is a maximum gain (5.15 dB) approximately twice (+3dB) that of the dipole antenna (2.15 dB). Although the half toroid radiation pattern is sufficient for DSRC communications, the increase in gain over the dipole antenna is not enough to compete with current DSRC antennas.

3.1.2 Loop Antennas

Figure 3-13: Isometric view of an ideal circular loop antenna, fed by a 50-Ω coaxial cable, modeled in Ansys-HFSS.

Yet another widely used wire antenna is the loop antenna. These antennas are relatively simple because they can be comprised of as little as a single wire loop connected to a transmission line. Along with monopoles and dipoles, loop antennas were among the first antennas ever developed for wireless communications. In fact, Heinrich

Hertz, whom the SI unit of frequency was named after, used loop antennas in his experiments in the 1800’s [ 21]. Although the circular loops Hertz used are the most common, loop antennas come in a wide variety of shapes and sizes. These antennas have received increased attention because many electronic circuits contain closed paths that can unintentionally act as loop antennas and cause undesirable radiation and interference within the circuit. An image of a circular loop antenna is shown in Figure 3-13.

For circular loop antenna’s such as the o ne depicted here, the parameter of most importance is the circumference of the loop. One of the most common circumferences in loop antenna design is the one wavelength loop antenna. One of the most well-known applications of loop antennas is as AM receivers for commercial radios. However, the circumference of these loops is much less the one wavelength, as a single wavelength in the AM band is around 100 meters.

In the DSRC band, on the other hand, the frequency is much higher than the AM band, and the wavelength that corresponds to this frequency is much shorter. This means one wavelength circumference loop antennas are of a more manageable size, as a wavelength is only about 50.8 mm long. Therefore, the antenna from Figure 3-16 was modeled as a one wavelength antenna with a center frequency of 5.8825 GHz, which corresponds to a circumference of only 50.8 mm. The reflection coefficient results from this are shown in Figure 3-14.

-5

-10

-15

-20

-25 4 4.5 5 5.5 6 6.5 7 7.5 8 Frequency (GHz) Figure 3-14: Reflection coefficient of an ideal, one-wavelength loop antenna fed by a 50- Ω coaxial cable , modeled in Ansys-HFSS, in decibels (dB).

This graph shows that loop antennas have a wide bandwidth compared to the other antennas reviewed. The dipole and monopole antennas both provided bandwidths of approximately 12.7%, while this loop antenna has a 10 dB bandwidth of 1 GHz, or

16.9 %. This bandwidth is a desirable characteristic for numerous applications.

However, in terms of DSRC, the extra bandwidth could be problematic. The DSRC band is only 75 MHz wide, and therefore, an antenna that has a bandwidth wider than 75 MHz will increase the amount of unwanted signals received. In this case, the lower cutoff frequency is below the wireless local area network band (5.15 – 5.35 GHz). This band is heavily utilized for in-home wireless networking, and it is a particularly important band to avoid overlapping. This is a major concern for implementing loop antennas for DSRC applications, and the other concern can be seen in Figure 3-15.

0° 0° 10 2.5 30° 330° 30° 330° 0 2 -10 1.5 60° -20 300° 60° 300° -30 1 -40 0.5 90° -50 270° 90° 0 270°

120° 240° 120° 240°

150° 210° 150° 210° 180° 180°

(a) (b) Figure 3-15: Radiation pattern in the (a) X-Y plane and (b) X-Z plane of an ideal, one- wavelength loop antenna fed by a 50-Ω coaxial cable , modeled in Ansys-HFSS, in decibels (dB).

The radiation pattern for the loop antenna reveals the omni-directional shape necessary for DSRC antennas, but it lacks the level of gain DSRC systems require. Like the dipole and monopole antennas previously examined, the loop antenna radiates at extreme elevation angles, and this lowers the gain it achieves at the lower elevation angles. This detriment, along with the wide bandwidth, prevents simple loop antennas from being applied to DSRC communications.

3.2 Common Variations on Fundamental Wire Antennas

Dipole, monopole, and loop antennas are the three most fundamental wire antennas. They have all been utilized in a number of different applications in the past, but they’ve all proven to have serious flaws when being applied to DSRC. However, these antennas were reviewed because many novel antenna designs use these fundamental variations as building blocks. The next three antennas are based on the previous antennas, but their performance is dramatically different. 38

3.2.1 Yagi-Uda Dipole Arrays and Folded Dipoles

Similar to dipole antennas, the Yagi-Uda antenna is one of the most popular antennas of the 20 th century. This antenna was first developed by Hidetsugu Yagi and

Shintaro Uda in 1927 in [23], but it wasn’t published in the United States in Engli sh until

1954 [24]. Since its initial introduction, it has gained widespread residential use as a means of receiving television signals transmitted over large distances. The basic shape of a Yagi-Uda antenna is shown in Figure 3-16.

Figure 3-16: Ansys-HFSS model of an ideal Yagi-Uda antenna array.

As Figure 3-16 shows, a Yagi-Uda antenna is an array of several antenna components. The basis of the antenna is the lone excited element, which in this case is a simple dipole antenna (located second from the left in Figure 3-16). On one side of the dipole antenna, a metal cylinder slightly longer than the dipole is positioned parallel to the dipole. This element is known as the reflector, named so because it inductively loads the dipole and reflects waves back towards antenna itself. A second reflector can be used to increase this effect and further decrease the antenna’s back lobe, but it is not necessary as it provides only marginal improvement over a single reflector design.

Positioned on the opposite side of the dipole as the reflector are a number of metal cylinders slightly shorter than the dipole antenna. These cylinders are referred to as 39 directors because they add a capacitive load to the dipole and act to draw electromagnetic waves away from the dipole and towards them. This effect, combined with the effect of the reflector, focuses the electromagnetic waves in a single direction. The amount that the waves are focused depends on the length and spacing of the reflectors and directors.

According to [22 ], the element should be spaced between 0.3λ – 0.4λ apart, and the director length should be between 0.4λ – 0.45λ. It goes on to state that the dipole length for a Yagi-Uda antenna should be between 0.45λ – 0.49λ, which is almost the same size as the dipole modeled and simulated previously (0.47λ – 0.48λ). All of this cumulates to create an extensive optimization process for developing this antenna.

0 4 4.5 5 5.5 6 6.5 7 7.5 8

Frequency (GHz) (a)

-2

-4

-6

-8

-10

-12

-14

-16 4 4.5 5 5.5 6 6.5 7 7.5 8 Frequency (GHz) (b) Figure 3-17: VSWR (a) and S 11, in decibels (dB), (b) of a Yagi-Uda antenna simulated in Ansys-HFSS.

Figure 3-17 shows the voltage standing wave ratio and reflection coefficient obtained when the ideal Yagi-Uda antenna from Figure 3-16 is simulated. These graphs reveal a couple of interesting characteristics of Yagi-Uda arrays. The important thing to note is the bandwidth difference between this type of antenna and the ideal dipole antenna analyzed previously. The dipole antenna had a single operation band and a single resonant frequency, whereas this antenna array has two different bands and two resonant frequencies. It is also important to notice that the two resonant frequencies are located very close to each other (< 0.5 GHz). The reason for this is the complex interactions between the dipole antenna at the center of the Yagi-Uda array and the directors around it. As the frequency increases past the resonant frequency, the dipole is becoming less well-matched, but because the directors are shorter than the dipole, they are becoming matched better. Therefore, the effect of the directors is to create a second

41 resonant frequency for the antenna as a whole. Also of interest is the fact that while there is a second resonant frequency, the entire operating bandwidth of the antenna is less than the bandwidth of the dipole antenna. Depending on the length, spacing, and number of directors used, the bandwidth and resonant frequency of a Yagi-Uda array can be altered to mitigate or amplify the effects mentioned above based on the application.

0° 0° 20 20 330° 30° 330° 30° 10 0 300° 0 60° 300° -20 60° -10 -40 270° -20 90° 270° -60 90°

240° 120° 240° 120°

210° 150° 210° 150° 180° 180°

(a) (b) Figure 3-18: Radiation pattern in the (a) X-Y plane and the (b) X-Z plane (dashed) and Y-Z plane (solid) of an ideal Yagi-Uda six-director antenna array simulated in Ansys- HFSS.

The characteristic that distinguishes a Yagi-Uda antenna array from the wire antennas previously analyzed is its highly directional radiation pattern. Where the monopole and dipole antennas radiate in all directions in the azimuth plane, the Yagi-Uda array only radiates in the direction the directors and reflectors are aligned in. This can clearly be seen above in Figure 3-18.

In addition to a highly directive radiation pattern, Figure 3-18 reveals significant side and back lobes. The back lobe can be controlled by altering the length and spacing of the reflector element, while the side lobes depend upon both the reflector and the

42 directors. They can be decreased or increased slightly as desired, but due to the omni directional nature of the dipole radiator at the center of the antenna, the side lobes cannot be removed completely. Also noteworthy is the observation that like the dipole and monopole antennas, the gain of the Yagi-Uda array decreases as the elevation angle moves away from the X-Y plane (θ = 90). Again, this can be attributed to the dipole radiator that forms the base of the Yagi-Uda antenna array. In this case, though, the directors are aligned with their centers lying on the X-Y plane. Therefore, as the elevation angle increase or decreases, the effect of the directors decreases drastically.

Because of this, the gain for the Yagi-Uda antenna decreases much more rapidly with elevation when compared to other wire antennas.

Although the highly directional radiation pattern of the Yagi-Uda antenna array makes it unsuitable for DSRC applications, it possesses one characteristic the fundamental wire antennas all lack. Where the maximum gain achievable from those antennas was still less than 6 dB, the six-director Yagi-Uda antenna shown in Figure 3-16 produces a maximum gain of over 12 dB. In fact, according to [22], with more directors and a finely tuned optimization process Yagi-Uda antennas can produce gains of over 17 dB. This is an excellent example of using parasitic elements to alter the operation of a wire antenna in order to adapt the antenna to different applications. The parasitic elements in this case are the reflector and the directors. They are referred to as

“parasitic” because they introduce an inductance (reflector) or capacitance (directors) to the antenna, even though they are not physically connected. Also, they can modify the radiation pattern of the antenna, but because they have no source, they cannot radiate

43 electromagnetic waves by themselves. This is a concept that will be revisited in depth later in this thesis.

Aside from the Yagi-Uda antenna, another variation on the dipole antenna is the folded dipole antenna. The folded dipole consists of a regular dipole antenna with the two ends connected to each other. Usually, the ends are connected with a wire with the same cross section as the dipole itself, but this is not always the case. An image of a folded dipole antenna can be seen in Figure 3-19 below.

Figure 3-19: Ansys – HFSS model of an ideal, half-wavlength folded dipole antenna.

This antenna typically uses the λ/2 version of the dipole antenna has a basis, which makes the full length of the antenna approximately equal to one wavelength. This is the equivalent of the circumference of the loop antenna analyzed previously. Because of this, the folded dipole antenna can be thought of as a modified dipole antenna, or a modified loop antenna. Neither definition is incorrect. However, it typically falls under its own classification because of its unique operation and widespread use.

As mentioned previously, the dipole antenna has an input of impedance of 75

Ohms, which makes it ideal for use with 50 Ohm transmission lines. The folded dipole antenna, on the other hand, has an input impedance of 300 Ohms (i.e. four times the dipole’s impedance), which makes it suitable for use with higher Z o transmission lines.

In addition to this, the reflection coefficient response for a typical folded dipole antenna differs from an equivalent loop antenna or dipole antenna. To illustrate this, the reflection coefficient responses of a dipole, folded dipole, and loop antenna have all been plotted together in Figure 3-20. In this case, all three antennas have been tuned to approximately the same center frequency (5.8825 GHz).

-5

-10

-15

-20

-25 4 4.5 5 5.5 6 6.5 7 7.5 8 Frequency (GHz)

Figure 3-20: Reflection coefficient response of a λ/2 dipole antenna (red, solid), a 1λ folded dipole antenna (blue, dashed) and a 1 λ loop antenna ( green, dotted) simulated in Ansys-HFSS, in decibels (dB).

While the shape of the folded dipole antenna appears to be a combination of a dipole antenna and a loop antenna, the reflection coefficient results in Figure 3-20 leave no doubt. The folded dipole has the same magnitude for the reflection coefficient at

5.8825 GHz as the dipole antenna (about -14.5 dB). However, the bandwidth for the folded dipole is larger than the bandwidth of the dipole antenna, but smaller than the bandwidth of the loop antenna. This is the primary advantage that folded dipole antennas have over half-wavelength dipole antennas. These antennas achieve a wider bandwidth than dipole antennas, but they don’t have the exceptionally large, potentially problematic bandwidths that loop antennas have.

0° 0° 3 5 330° 30° 330° 30° 2.5 0 2 -5 -10 300° 1.5 60° 300° 60° -15 1 -20 0.5 -25 270° 0 90° 270° -30 90°

240° 120° 240° 120°

210° 150° 210° 150° 180° 180°

(a) (b)

Figure 3-21: Radiation pattern in the (a) X-Y plane and (b) X-Z plane (dashed) and Y-Z plane (solid) of an ideal, one-wavelength folded dipole antenna, modeled in Ansys-HFSS.

Unfortunately, because folded dipoles are similar to loop antennas and dipole antennas, they also suffer from the low gain that the other two antenna types share. This can be seen in the radiation pattern in Figure 3-21. Also, because the folded dipole is not symmetrical about the Z axis, its radiation pattern is not symmetrical either. As stated in

Chapter 1, an omni directional radiation pattern is essential for DSRC applications.

Folded dipoles also present the same free-space mounting challenges that make dipole difficult to implement in automobiles. Because of these drawbacks, folded dipole antennas are not suitable for future DSRC communications.

3.2.2 Helical Antennas

A helix antenna is another common wire antenna variation. This antenna consists of a wire coil mounted vertically, usually on an ideal ground plane. An image of this can be seen in Figure 3-22 (a). Unlike the fundamental wire antennas, there are a large number of variables that affect the performance of helical antennas. This leads to both a complicated design process and a wide variety of applications for this type of antenna. In order to emphasize this point, some of the variable that dictate helical antennas have been called out in Figure 3-22 (b) and explained in Table 3-1. Also unlike the previous wire antennas, helical antennas can radiate electromagnetic waves through two different radiating methods, which makes them incredibly versatile antennas.

A H

(a) (b)

Figure 3-22: 3-dimensional view (a) and 2-dimensional view (b) of a helical wire antenna, modeled in Ansys-HFSS

Table 3.1: Variables affecting helical antenna performance.

Symbol Description Symbol Description

RH radius of the helix Α pitch angle RW wire radius H antenna height C circumference of the helix L length of one revolution S spacing between helix turns N number of revolutions

The two modes that a helix antenna can operate in are known as axial mode and normal mode. According to [25], the mode the antenna operates in depends on the overall length of the wire, which can be calculated by multiplying the number of revolutions by the length of one revolution (i.e. nL from Table 3.1). If this length is greater than one wavelength, the antenna radiates in axial mode. If this length is much less than one wavelength, it radiates in normal mode.

When the antenna radiates in axial mode, it produces a beam radiation pattern similar to the Yagi-Uda antenna already discussed. However, because of the shape of the helix, it radiates with circular polarization rather than the vertical polarization necessary for DSRC. By varying the radius of the helix to make a hemispherical helix shape, the antenna can be made to radiate over a wider range of angles [26]. Also, there are many axial mode variations that can be made simple by adjusting the parameters in Table 3-1.

If the pitch angle, α, and the spacing between turns, S, are set to zero, and the helix radius, R, increases with subsequent turns, the result is a 2-dimensional spiral antenna.

This type of antenna can be fed with a beam antenna, or even a second helix antenna

[27]. However, this type of antenna also has circular, not vertical, polarization.

Additionally, while both of these axial mode antennas are improvements over the basic helical antenna, they still do not radiate at low angles as is required for DSRC. For this reason, axial mode radiation is inappropriate for this application.

Normal mode radiation occurs when the dimensions of the antenna are electrically very small (i.e. nL << λ). In this mode, the antenna is essentially a combination of small dipoles and small loop antennas connected in series [22]. Because of this, the radiation pattern is omni directional in the same way the loop antenna and dipole antenna are omni directional. These normal mode helical antennas are preferred for some automobile applications, like the tire pressure monitoring system [28], because they offer a significant slight reduction over classic monopole antennas.

This benefit can readily be seen in the folded spherical helical antenna presented in [29]. To create this antenna from the parameters listed in Table 3-1, the radius of the helix, R, is decreased on subsequent turns while the pitch angle, α, and the spacing between turns, S, remain constant. This particular design utilizes four such helical antennas connected to each other at the top of the antenna in order to create a hemisphere shape. As previously mentioned, antennas of this shape radiate axially and produce a wide beam antenna [26]. If the hemisphere dimensions are small, though, this design can alternatively be made to radiate in the normal mode instead [29].

While these helical antennas offer exceptional performance in the TPMS range

(300 – 400 MHz), the size reduction benefit they provide at these frequencies becomes a deterrent for DSRC antenna design. As previously mentioned, the length and overall height of helical antennas must be much less than one wavelength in order to achieve normal mode operation (nL << λ). In [2 9], the overall height of the four-arm hemisphere helical antenna designed is 8.25 centimeters at a center frequency of 210 MHz. This is equal to 0.0578λ. At t he center frequency for DSRC communications, 5.8825 GHz, the nominal value of 0.0578λ becomes 2.94 millimeters. In addition to this, in order to

49 achieve the same performance at this higher frequency, some of the other parameters would need to be reduced as well. At such a small size, the level of detail and the intricacy involved in manufacturing the antenna is prohibitive. Furthermore, after the antenna is constructed, the fine wire would need a support structure in order to withstand the vibrations and abuse that vehicular components are subjected to. These difficulties make helical antennas a poor choice for DSRC communications.

Up to this point, many wire antennas have been analyzed for DSRC communications. Dipoles, monopoles, and loop antennas are fundamental wire antenna designs that are all insufficient for this new application. Yagi-Uda arrays, folded dipoles, and helical antennas are slight variations of the classic antenna designs that like fall short of the necessary performance specifications. Because of this, there is an opportunity for a novel antenna design that successfully meets or exceeds the predetermined requirements for DSRC communications, with a center frequency of 5.8825 GHz. The novel antenna presented and analyzed in this thesis meets these criteria and provides a practical solution for the future of DSRC communications in vehicles.

Chapter 4

Novel DSRC Antenna Design and Simulation

4.1 Novel Antenna Design

With the need for a novel DSRC antenna design clearly defined, the design process focuses on enhancing the strengths of previous wire antenna designs while mitigating their weaknesses. The shortcomings of the antennas already discussed should be clear at this point. None of the antennas analyzed provides a high gain, easily manufactured, 5.8825 GHz antenna with vertical polarization and an omni directional radiation pattern at low elevation angles. However, all of the antennas fulfill some of these requirements.

All of the antennas presented can achieve vertical polarization. The helical antenna, though, must radiate in normal mode for this, and at 5.8825 GHz, normal mode helical antennas are too small and fragile to be manufactured for vehicle use. Although loop antennas can also radiate with vertical polarization, they tend to have a much wider bandwidth than other wire antennas. This can introduce unwanted interference from the nearby WLAN frequency band. Therefore, the weaknesses of helical and loop antennas

51 are severe, and extremely difficult to overcome. This makes them unattractive as a base for novel antenna designs.

This leaves dipoles, monopoles, Yagi-Uda arrays, and folded dipoles as potential bases for a novel antenna design. Of these four antennas, monopoles are the only antenna type that is well suited for vehicular applications. This is because, as previously stated, monopole antennas need to be mounted on electrically large ground planes where the other antennas all radiate in free space, and any DSRC antenna will most likely be mounted on the roof of the vehicle, which acts like a large ground plane. Therefore, of the wire antennas presented, monopole antennas provide the best starting point.

In the analysis of monopole antennas, it was determined that monopoles lack adequate gain at low elevation angles for DSRC communications. Most of the other antennas presented also suffered from this same shortcoming. The only exception to this was the Yagi-Uda array antenna, which could be used to achieve incredibly high gains

(~17 dB). However, the Yagi-Uda array only achieves this gain in one direction via the use of reflector and director elements. For this novel antenna design, the omni directional radiation pattern of the monopole antenna is combined with the increased gain of the

Yagi-Uda array to create a new antenna that meets all the specifications for DSRC.

During the analysis of the Yagi-Uda array, it was determined that there is only one radiating element in an array, which is typically a λ/2 dipole antenna. However,

Yagi-Uda antennas have been designed that use other antennas for the radiating element, including microstrip patch antennas [30] and monopole antennas [31]. For the monopole variation, the λ/2 dipole is replaced with a λ/4 monopole, and the directors are likewise shortened. The distances between the elements of the array, however, remains

52 unchanged. In the Yagi-Uda array discussion, it was noted that the actual length of the active dipole should be between 0.45λ – 0.49λ, and director length should be between

0.4 λ – 0.45λ. When halved for monopole design, this corresponds to a range of 0.225 λ –

0.245λ for the monopole, 0.2λ – 0.225λ for the directors, and slightly longer than λ /4 for the reflector. The entire theoretical range of values for the parameters of a monopole

Yagi-Uda antenna with a target frequency of 5.8825 GHz can be seen in Table 4.1.

Table 4.1: Parameters of a monopole Yagi-Uda antenna array.

Symbol Description Electrical Range Nominal Range

LD Length of driven element 0.225λ – 0.245λ 11.4 mm. – 12.5 mm. LE Length of directors 0.2λ – 0.225λ 10.2 mm. – 11.4 mm. LR Length of reflector >0.25 λ >12.7 mm. S Spacing between elements 0.3λ – 0.4λ 15.3 mm. – 20.3 mm. N Number of directors n/a ≥1

The values in Table 4.1 can be used to design an antenna that successfully combines monopole antennas and Yagi-Uda arrays. Unfortunately, this type of antenna is still highly directional, and therefore, more alterations are needed to make it suitable for DSRC. In order to accomplish this, the gain of the antenna needs to be increased in all directions in the X-Y plane, rather than only one.

Figure 4-1: An ideal model of a novel, compact, wire antenna array on an infinte ground plane, designed for mobile units used in future DSRC communications.

In the Yagi-Uda antenna analysis, it was noted that the gain of the array is decreased in one direction and increased in the other via the inductive loading of the reflector element and the capacitive loading of the director elements, respectively. Since the gain of a DSRC antenna does not need to be diminished in any direction in they X-Y plane, the reflector can be removed from the antenna entirely. After the reflector is removed, the directors can be repositioned around the central monopole so the radiation pattern is not focused in one specific direction. The directors should be positioned around the monopole in a circle with a radius determined by the value denoted by S in

Table 4-1. This has the desired affect of increasing the gain in the X-Y plane at the expense of the gain at higher elevation angles. The resulting novel antenna design satisfies all the requirements for DSRC communication. An ideal model of this antenna can be seen in Figure 4-1. This novel design will be analyzed and simulated later in this chapter.

The performance of the novel antenna array presented in the previous chapter is governed by many parameters. All of the design variables that were considered for the

Yagi-Uda array also need to be considered for this novel antenna. This increases the complexity of an already complex design process. In order to illustrate this, a model of the ideal monopole array is depicted in Figure 4-2 with the design variables called out.

The effects of these variables and the performance of the antenna will be simulated and analyzed in this chapter. Also, a feed board used to integrate the antenna into an automobile roof is presented, and the performance of the antenna with the feed board attached is simulated and analyzed.

Figure 4-2: An ideal model of a wire antenna array shown with the design variables that affect the antenna performance.

4.2 Antenna Simulation

The most important design variable for this antenna is the number of director elements used, or N from Table 4-1. It is considered the most important because the directors are responsible for creating the antenna’s omni directional radiation pattern. An acceptable antenna design utilizes enough directors to achieve the desired effect, but not so many directors that material resources are wasted. Therefore, the number of directors was varied, and the radiation pattern results are shown in Figure 4-3. For this parametric sweep, the value of the other design variables is not crucial because they primarily affect the magnitude of the radiation pattern, while the number of directors affects the overall shape of the radiation pattern. The other parameters were simply chosen to be within the ranges specified in Table 4-1.

0° 10 30° 330° 8

6 60° 300° 3 Elements 4 4 Elements 2 5 Elements 90° 0 270° 6 Elements 7 Elements 8 Elements 120° 240°

150° 210°

180°

Figure 4-3: X-Y plane gain patterns of a novel wire antenna array with a varying number of parasitic elements, measured in dB at 5.9 GHz.

This graph reveals drastic changes in the shape of the radiation pattern for antenna arrays with different numbers of evenly spaced directors. The three director array achieves a positive gain for every direction in the X-Y plane, however there is a large magnitude difference (~3 dB) between the maximum and minium gains. For strong signals, this wouldn’t be an issue because the gain at the minimums would be sufficient.

For weak signals, though, the three director array would only operate properly if the direction of maximum gain were pointed towards the signal. This is undesirable as the signal direction in DSRC communications is time-varying and highly stochastic.

Figure 4-3 also reveals that, as predicted, each additional director makes the raditation pattern more omni directional. When comparing the three director radiation pattern to the four director pattern, though, it is evident that the minimums in the four direction pattern are lower than those in the three director pattern. This phenomenon can

56 be explained by the lack of reflectors. In th Yagi-Uda antenna array, a reflector element is used because the director elements create a sizable back lobe located 180 degrees from the main lobe. In the three element array, the main lobes are spaced 120 degrees apart, located at 90, 210, and 330 degrees, and the corresponding back lobes are 180 degrees away from the main lobes at 30, 150, and 270 degrees. These back lobes coincide with the angles farthest from the directors, where a gain minimum would be located.

On the other hand, for the four director antenna array, the direcors are spaced 90 degrees apart rather than 120 degrees. Because of this, the back lobes of the directors, which are located 180 degrees away from the directors, always occur in the same location as one of the other directors. In short, the back lobe from one director is reinforcing the main lobe of the opposite director. This arrangement makes the overall gain of the antenna higher, but it also makes the radiation pattern less omni directional.

The increased overall gain that the antenna arrays with even amounts of directors provide makes the eight director array the best choice in this case. Although the radiation pattern is almost perfectly omni directional for the seven element array, adding the eighth element maintains omni directionality while incresing the average gain from 6.62 dB to

6.72 dB. Furthermore, adding additional directors does not further improve an already omni directional radiation pattern, and the average gain increases from 6.72 dB to only

6.77 dB for the ninth director and 6.84 dB for the tenth. This limited benefit makes the eight director array more attractive.

After the number of directors is decided, the next pair of parameters to consider is the length of the monople and the directors. These two parameters are considered together because they are dependent on each other. The directors must be slightly shorter

57 than the monopole, or the will act as reflectors instead of directors. For this ideal design, it was determined through simulation that a length of 0.21λ (10.6 mm.) for the monopole and 0.20λ (10.2 mm) for the director provided the best solution. This puts the director length at the short end of the theoretical range for Yagi-Uda antenna design, and the monopole is actually too short for an ideal Yagi-Uda antenna at 5.9 GHz. Although since this design is significantly different from a Yagi-Uda antenna, some differences in the design process should be expected.

The final parameter important for this design is the spacing of the directors. In the Yagi-Uda antenna, the director spacing, S, defines the length from one director to the next. For this design, the directors are located on a circle with the driven element at the center, and they are separated from each other angularly by 45 degrees. Therefore, the variable S in this case refers to the radius of the circle of directors positioned around the monopole. In the analysis of Yagi-Uda antennas, it was determined that the theoretically perfect spacing value of λ/2 needed to be shortened for practical applications.

Furthermore, Table 4-1 specifed 0.3λ – 0.4λ as the acceptable range for the director spacing. Since this design is based on the Yagi-Uda antenna, the circle radius, S, was varied using the mid point, 0.35λ (17.8 mm) as the st arting point. The results of this parametric sweep revealed that the director spacing does not significantly affect the reflection coefficent, nor the shape of the radiation pattern in the X-Y plane. However, the spacing does have a major affect on the radiation pattern in the Y-Z plane. This can clearly be seen in Figure 4-4.

0° 7 30° 330° 6 5

60° 4 300° 3 2 1 S = 0.35 λ 90° 270° 0 S = 0.45 λ S = 0.55 λ

120° 240°

150° 210°

180°

Figure 4-4: X-Z plane gain patterns of a novel wire antenna array with varying spacing between the monopole and parasitic elements, measured in dB at 5.9 GHz.

As this plot shows, the director spacing affects the maximum gain of this wire antenna array. This is as expected, because altering the director spacing changes affects half-wavelength radiating mechanism. This mechanism is the reason the directors act to increase the gain of the Yagi-Uda antenna, and it is the reason the director increase the gain of this antenna as well. If the antennas are not spaced properly, they will not resonate with the center monopole and reradiate incident electormagnetic waves.

Figure 4-4 also shows that 0.45λ (22.9 mm) is the optimum director spacing.

Again, this disagrees with the Yagi-Uda antenna where the range for S is given to be 0.3λ

– 0.4λ. Although, this is still less than the theoretical wavelength of λ/2. This discrepancy can be attributed to one of the significant differences in design. The Yagi-

Uda has multiple directors aligned one right after the other. Therefore, each director is

59 affected by both the director before it and the one right after it. The novel antenna presented in this thesis, though, has only a single ring of directors. Since each director is coupled only to the driven monopole in the center, the proper spacing is closer to the ideal value of λ/2.

Table 4.2: Optimized Parameters of a novel wire antenna array.

Symbol Description Electrical Value Nominal Value

LD Length of driven element 0.21λ 10.6 mm. LE Length of directors 0.2λ 10.2 mm. Spacing between S 0.35λ 22.9 mm. elements N Number of directors n/a 8 D Diameter of Wires 0.008λ 0.4 mm.

After the director spacing is determined, the last parameter is the diamter of the wires used for the monopole antenna and the director elements. Since the monopole utilizes the same radiation mechanism as the dipole antenna, the radius sweep shown in

Figure 3-8 applies to this antenna as well. Therefore, the radius of the antenna has a slight affect on the bandwidth of the antenna but not the center frequency. Based on this information, a dipole radius of 0.2 mm was chosen for this design. The parameters for the ideal novel antenna array designed for a center frequency of 5.9 GHz are shown above in

Table 4.2, and the performance results are shown in Figure 4-5 and Figure 4-6.

-5

-10

-15

-20

-25

-30

-35 4 4.5 5 5.5 6 6.5 7 7.5 8

Figure 4-5: Reflection coefficient, S 11 , of an ideal model of a novel, compact, wire antenna array on an infinte ground plane simulated in HFSS (blue, solid) and CST (red, dashed), in decibels (dB).

0° 0° 8 20 330° 30° 330° 30° 6 0 300° 4 60° 300° -20 60° 2 -40 270° 0 90° 270° -60 90°

240° 120° 240° 120°

210° 150° 210° 150° 180° 180°

(a) (b) Figure 4-6: Radiation pattern in the (a) X-Y plane and (b) X-Z plane of an ideal model of a novel, compact, wire antenna array on an infinte ground plane, in decibels (dB).

4.3 Feed Board Design

As previously stated, the input impedance of a dipole antenna is approximately 75

Ohms. Since a monopole antenna is one-half of a dipole antenna, the input impedance of a monopole antenna is therefore half the input impedance of a dipole antenna, which ends

61 up being exactly 36.8 Ohms. However, the characteristic impedance of most transmission lines is 50 Ohms. This impedance mismatch will decrease the amount of power delivered to the monopole, which will in turn decrease the overall performance of the antenna.

A second obstacle for this antenna is the feed point orientation of the monopole.

The wire array must be mounted vertically in order to achieve the necessary omni directional radiation pattern. Since the active element is the monopole in the center of the array in Figure 4-1, this would imply an SMA connector connecting directly to the center of the antenna. This is impractical, because connecting the coaxial input cable to the antenna would add unwanted thickness to the antenna design. An elbow SMA connector can be used to mitigate this slightly, but it would not alleviate the problem altogether.

To both reduce the effect of the impedance mismatch and provide a low-profile means of connecting to the antenna, a feed board for the monopole array is presented.

This board uses a grounded coplanar waveguide (GCPW) transmission line to connect a side-mounted SMA connector to the monopole antenna in the center of the board. This type of transmission line provides both the ground plane necessary for the monopole array and the coplanar ground plane needed for the SMA connector. In addition, the impedance mismatch is mitigated by designing a quarter-wave transformer into the grounded coplanar waveguide transmission line. The monopole antenna at the center of the array is connected to one end of the transformer, while the other end is connected to a section of 50 Ohm GCPW. This 50 Ohm section carries the signal from the transformer to the SMA connector. An image of the entire feed board is shown in Figure 4-7.

(a) (b)

(c) Figure 4-7: Model of a feed board for a monopole array with the top view (a), bottom view (b), and transformer detailed view (c) shown.

The top of the feed board shown in Figure 4-7 (a) shows a small circle cut out of the ground plane. This is necessary for the monopole to connect to the feeding line through the substrate without coming into contact with the ground plane. This allows the monopole antenna to be vertically oriented through the substrate, and it provides a large ground plane for the directors to be mounted on. The size of this feed board was strategically designed to be 55 mm by 55 mm in order to fit the entire circle of directors, which has a diameter of 45.8 mm. The substrate used in this design is FR-4 Epoxy with a dielectric constant of 6.15, which was chosen for its accessibility.

The transmission line and quarter-wave transformer shown in Figure 4-7 (c) are the important design features of the feed board. The transmission line should have a characteristic impedance of 50 Ohms to match the transmission line input from the vehicle. The quarter-wave transformer, on the other hand, should have a characteristic impedance between the characteristic impedance of the feed line (50 Ohms) and the input impedance of the monopole antenna (36.8 Ohms). The exact value of this impedance can be calculated using equation 2 [32].

Z = Z R 1 o L (2)

In equation (2), Z 1 is the characteristic impedance of the quarter-wave transformer, Z O is the characteristic impedance of the feed line, and R L is the input impedance of the monopole antenna. Substituting the values specified above into equation (2) results in a characteristic impedance of 43 Ohms for the quarter-wave transformer. Therefore, the grounded coplanar waveguide transmission line must transition from a 50 Ohm characteristic impedance to a 43 Ohm characteristic impedance to properly match the antenna and the input cable.

The characteristic impedance of a transmission line is dependent on several parameters. A cross section of a GCPW transmission line with these parameters called out is shown here in Figure 4-8. In this figure, Ts is the thickness of the substrate, Tc is the thickness of the conductor, Wt is the width of the trace, and Wc is the width of the entire section cut out of the conductor.

Wt εo

εrεo Ts

Figure 4-8: Cross section of a grounded coplanar waveguide (GCPW) transmission line.

This image shows the many different variables that are used to calculate the characteristic impedance of a GCPW transmission line. Figure 4-8 also reveals a unique characteristic of this type of line. Unlike other transmission lines, the electromagnetic wave in a GCPW line travels between regions with two different dielectric constants, because a portion of the wave is located above the main trace while the rest is located below it. Above the main trace is free space, with a dielectric constant of one. Below the main trace is the substrate with a dielectric constant of 6.15. In this case, the electromagnetic wave traveling along the trace is coupled to both the coplanar ground plane and the ground plane on the opposite side of the substrate. Because of this, the portion of the wave that travels through the substrate depends upon how strongly the wave is coupled to each ground plane. Therefore, neither the dielectric constant of air nor the dielectric constant of the substrate accurately reflects the dielectric constant of the transmission line as a whole. The effective dielectric constant, ε eff , can instead be calculated using equation (3). In this equation, the value of k is the ratio of the trace width over the cutout width of the transmission line, and the value of k1 is based on the ratio of these same values to the thickness of the substrate [33].

+ e K(k )' K(k1) 0.1 r e = K(k) K(k1 )' eff + K(k )' K(k1) 0.1 (3) K(k) K(k1 )'

In equation (3), the K function is known as the complete elliptical integral of the first kind. This is a recursive function that requires several iterations to solve. A full explanation of this equation is presented in [33]. For this thesis, when this equation is solved, the value is ε eff can be entered in equation (4) in order to find the characteristic impedance of the GCPW transmission line.

60 0. p 0.1 Z = o e K(k) K(k ) eff + 1 (4) K(k )' K(k1 )'

Utilizing these equations, it is possible to determine the characteristic impedance of a transmission line given the parameters shown in Figure 4-8. Of the parameters in that figure, though, the substrate thickness, conductor thickness, and substrate dielectric constant are all constant for the entire board. This leaves varying the trace width and the width of the cut out as the only means of altering the transmission line’s characteristic impedance to create the quarter-wavelength transmission line. Furthermore, to simplify the manufacturing process and the design process, it is desirable to keep one of these values constant. As Figure 4-7 (c) shows, for this design the cut out width was held constant at 3.4 mm, and the trace width was varied. Using this approach, a 50 Ohm

GCPW transmission line was obtained when the trace width was set to 1.8 mm, and a 43

Ohm transmission line was achieved by setting the trace width to 2.2 mm. With the widths of the transmission line determined, the last variable to consider is the length.

The length of the transformer should be λ/4, denoted by the given name of quarter-wave transformer. This length will cause the components of travelling waves reflected at the two impedance discontinuities to sum to zero. This greatly reduces the insertion loss that occurs when sources and loads are mismatched. However, for this benefit to be realized, the transformer section must be exactly λ/4 at 5.9 GHz.

In the previous discussions of wavelength, the wavelength could be determined by dividing the speed of light in a vacuum by the desired frequency. This equation is accurate because electromagnetic waves travel at the speed of light in free space, and the antenna design discussions all dealt with antennas radiating electromagnetic waves into free space. Since the dielectric constant of the substrate is greater than the dielectric constant of free space, it slows down the travelling wave to slightly less than the speed of light in a vacuum, which makes the actual wavelength shorter. Therefore, in order to factor in the affect of the substrate, λ/4 should be found using equation ( 5). l = c e 4 4 fo eff (5)

In this equation, c is the speed of light in a vacuum, f o is the center frequency, and

εeff is the effective dielectric constant of the substrate found in equation (3). Since the effective dielectric constant is dependent on the trace width and the cut out width, the wavelength of the signal along the transmission line is also dependent on those two values. Therefore, equation 5 must be solved using the effective dielectric constant of the quarter wave transformer section of the transmission line. In this case, λ/4 at 5.9 GHz is found to be 6.3 mm. Therefore, the length of the transformer is set to this value, and the design process for the feed board is complete.

TABLE 4.3: Feed Network Board Details Symbol Parameter Value n/a Board Dimension 55 mm × 55 mm εr Substrate Dielectric Constant 6.15 Ts Substrate Thickness 1.524 mm Tc Copper Thickness 0.035 mm W50 50 Ω T ransmission line Trace Width 1.8 mm G50 50 Ω T ransmission line Gap Width 0.8 mm L50 50 Ω T ransmission line Length 15 mm W43 Impedance Transformer Trace Width 2.2 mm G43 Impedance Transformer Gap Width 0.6 mm L43 Impedance Transformer Length 6.3 mm

In order to determine the effectiveness of the presented transformer design, the board was simulated without the antenna array in Ansys HFSS. This simulation provides the reflection coefficient and transmission coefficient. Ideally, the transmission coefficient, S21, would be equal to zero, as this would indicate that the entire electromagnetic wave that enters port 1 of the transmission line exits port 2. Also, the reflection coefficient, S11, should be low to ensure that none of the wave that enters port

1 is reflected back towards port one. Figure 4-9 below shows that this board accomplishes these two goals.

-5

-10

-15

-20

-25

-30 4 4.5 5 5.5 6 6.5 7 7.5 8

Figure 4-9: Transmission coefficient (red, dashed) and reflection coefficient (blue, solid) results of the quarter wave transformer, simulated in Ansys-HFSS, in decibels (dB).

4.4 Cavity Design

This board can be combined with the antenna array designed in section 4.1 to provide a viable antenna for DSRC communications. However, there are still a few minor flaws in the design that need to be addressed before it could be mounted in a vehicle. First, the feed board presented is a double sided board. This was specifically chosen to provide ground planes on both sides of the board so that the antenna array and input cable could both be connected easily. The flaw with this design is that double sided boards are more prone to exciting parallel plate modes between the two conductors. This occurs when feed board acts as a parallel plate waveguide, and electromagnetic waves travel through the substrate. This situation is undesirable because parallel plate modes in this board would be difficult to predict and impossible to control. Therefore, they could cause stochastic interference, and they could detrimentally affect the antenna radiation if left unchecked. The simplest solution to mitigate unwanted parallel plate modes is to connect the two ground planes with vias, which are holes through the substrate where the two sides of the board can be soldered together. Shorting the two ground planes together in such manner effectively grounds traveling electromagnetic waves in the substrate.

The second imperfection in this design is the lack of a clear mounting strategy.

While the board could certainly be attached to the roof of a vehicle, the printed circuit board lacks the structural integrity necessary for vehicular environments. Without an external support, the strength and durability of this design is insufficient. Unfortunately, any metallic mounting structure will affect the radiation characteristics of the antenna system because it is in the near field region of the antenna. In order to account for this,

69 the mounting structure should be simulated along with the antenna to determine the effect it will have and mitigate any detrimental side effects. The last concern this design fails to address is the potential for interference from wireless communication systems within the vehicle. On modern vehicles, there are multiple wireless systems radiating at many different frequencies as discussed in chapter one. These include remote keyless entry systems, tire pressure monitoring systems, and many others. In addition to these systems, vehicle passengers routinely introduce a variety of their own wireless electronic devices. These devices could include Bluetooth equipped electronics, mobile phones, portable video gaming devices, and medical equipment (i.e. insulin pumps, pacemakers, etc.) as well as many others. If the feed board presented above were integrated into the roof of the vehicle, the bottom feed trace and quarter-wave transformer would be exposed to electromagnetic waves from all of these devices and any other wireless systems in the vehicle.

Fortunately, all three of these flaws can be corrected with two simple additions.

Adding via holes and a metallic cavity to the back of the feed board is a simple, economical solution to these issues. The metal cavity provides the structure the system needs to withstand the extreme environment of a typical vehicle. Also, the metal doubles as an electromagnetic shield for the underside of the feed board. This effectively prevents nuisance radiation from passing through the metal and affecting the signal along the feed line. Therefore, the presence of the metal alone relieves two of the three major flaws in this design.

The last potential problem with this design is the tendency for the design to excite parallel plate modes. As previously mentioned, vias are used to eliminate parallel plate

70 modes by connecting the two ground planes together around the transmission line. Also, this designs uses strategically placed screws around the perimeter of the board to attach the cavity to the feed board. These screws serve a dual purpose. First, they provide a durable and cost effective means of mounting the board to the cavity. Secondly, the metallic screws electrically connect the two ground planes together, producing the same effect around the perimeter of the board as the vias produce around the transmission line.

This connection suppresses parallel plate modes and thereby increases the overall stability of the system. An image of the feed board with the additional vias can be seen in Figure 4-10, while the cavity is shown in Figure 4-11.

(a) (b)

Figure 4-10: Feed board (a) top and (b) bottom with vias added around transmission line and screws added around perimeter.

Figure 4-11: Metallic cavity design to protect and shield GCPW feed line.

4.5 System Assembly and Simulation

The last step in assessing this novel antenna design’s potential for DSRC applications is to simulate the entire assembled antenna system. The novel array dimensions were then tuned to correct for the effects of the feed board and cavity. This resulted in a final monopole length of 11.6 mm, a director length of 11.3 mm, and the director spacing, 22.9 mm, was not changed. The system, consisting of the antenna array, feed board, cavity, and mounting screws, can be seen in its entirety in Figure 4-12 (a) and

(b). Figure 4-13 depicts the reflection coefficient, S 11 , of the assembled system compared to the S 11 response of the antenna by itself. Finally, Figure 4-14 shows the radiation pattern of the antenna system.

(a)

(b) Figure 4-12: Complete, novel antenna array system with antenna array, feed board, and cavity, shown in (a) exploded view and (b) assembled view. 0 10

-5 8

-10 6

-15 4

-20 2

-25 0 4 4.5 5 5.5 6 6.5 7 7.5 8 4 4.5 5 5.5 6 6.5 7 7.5 8

Figure 4-13: (a) Reflection coefficient, S 11 , and (b) VSWR of the fully assembled, novel antenna system, in decibels (dB), with respect to frequency (GHz). 0° 0° 10 10 330° 30° 330° 30° 8 5 6 0 300° 60° 300° -5 60° 4 -10 2 -15 270° 0 90° 270° -20 90°

240° 120° 240° 120°

210° 150° 210° 150° 180° 180°

(a) (b) Figure 4-14: Radiation pattern in the (a) X-Y plane and (b) X-Z plane of the assembled, novel antenna array system, simulated in Ansy-HFSS, in decibels (dB).

These results show that the novel antenna system presented in this thesis meets all the requirements for vehicle-to-vehicle and vehicle-to-infrastructure communications in the DSRC frequency band. The reflection coefficient response in Figure 4-13 shows a strong resonance at 5.8825 GHz, and a -10 dB bandwidth of 400 MHz covering 5.6 – 6.0

GHz, which includes the entire DSRC band. The radiation pattern in Figure 4-14 shows omni directionality in the X-Y plane. The gain achieved in this plane is over 7 dB, which is higher than any of current DSRC antennas researched. Furthermore, the total height of this antenna system is approximately 20 mm. This low profile is ideal for concealing the antenna in the roof of the vehicle to minimize the antenna’s visual impact on the vehicle.

Chapter 5

Conclusion and Future Work

5.1 Conclusions

In this thesis, simulation results are presented for a novel wire antenna array for use in future Dedicated Short Range Communications. The design is modelled and simulated using industry standard Anys-HFSS software, and it is further validated by confirming these results utilizing CST Microwave Studio Suite software.

Along with the novel antenna design, a feed board is proposed to improve the impedance match between the novel antenna array and a standard 50 Ohm transmission line. This feed board has the desirable characteristics of a low profile, acceptable impedance match, and easy element mounting through the use of a double sided substrate. A metallic cavity along with mounting screws is attached to the board to improve the durability of the design. Also, the cavity has the added benefit of shielding the antenna array from electromagnetic waves within the vehicle while suppressing parallel plate modes within the feed board substrate. In addition to this, vias were added to further suppress parallel plate modes and increase the overall stability of the entire antenna array system.

Finally, the entire system is assembled and analyzed. The results from these simulations show an adequate -10 dB bandwidth in the S 11 response of the antenna system. The radiation pattern of the system reveals an omni directional antenna ideal for

DSRC applications. At low angles, this novel system provides a significant gain improvement over current DSRC antennas. Additionally, this system maintains a low physical profile that would allow it to be integrated into a vehicle’s roof without cosmetically altering the vehicle.

5.2 Future Work

Although this thesis presents the design and evaluation of a novel antenna system for DSRC, there is still more work to be done in this area. In order to be considered for use in vehicles, this antenna would need to first be fabricated, and the physical antenna would need to be thoroughly tested in an anechoic chamber. Through real-world testing, this antenna will undoubtedly require slight tuning to achieve the theoretical performance presented in this thesis. The proper method for tuning this antenna is to first manufacture the board with a monopole and directors slightly longer (i.e. 1 – 2 mm excess) than the optimized length. After achieving the initial performance parameters, the antenna elements may be trimmed as much as necessary. This is the preferred approach, as it is much more effective than attempting to add metal to a radiating element that is too short.

After the individual antenna array system is tested and tuned, it will have to be integrated into an actual vehicle and evaluated again. This can be a lengthy and expensive process as there are many government regulations that must be passed for

76 antennas used in vehicles. In addition to this, renting/purchasing the equipment needed to perform these full-vehicle evaluations requires more resources. Also, these tests must be repeated for every vehicle the antenna is applied to. Because of this financial burden, this evaluation is typically left for the vehicle manufacturer to conduct. Also, antennas are often evaluated in vehicles before they have been put in production. Therefore, vehicle manufactures typically prefer to do this testing with their own facilities and employees to protect confidential prototype vehicles.

While the aforementioned work includes many required evaluations that cannot be overlooked, it is reasonable to expect this antenna to pass these tests with only minimal adjustment. The reason for this is the monopole antenna at the heart of this design. As mentioned in Chapter 3, the monopole antenna has been thoroughly analyzed and is, in general, a well-understood antenna. This provides confidence that the tuning process will proceed as mentioned, and the final result will be satisfactory as predicted.

In conclusion, while this novel antenna array may need to be slightly altered and tuned, fundamentally the design presented in this thesis can be directly applied to future

Dedicated Short Range Communications to increase driver convenience and more importantly, driver safety.

References

1. IEEE Standard for Information Techonology - Telecommunications and information

exchange between systems - Local and Metropolitan Area Networks - Specific

Requirements, Amemdment 6: Wireless Access in Vehicular Environments, 2010.

2. FCC Report and Order, ET Docket No. 98-95, RM-9096, October 22, 1999.

3. G. Karagiannis et al., "Vehicular networking: a survey and tutorial on requirements,

architectures, challenges, standards and solutions," IEEE Communications Surveys &

Tutorials , vol. 13, no. 4, pp. 584-616, July 2011.

4. B. Cronin, “Intellidrive Program Overview,” U.S. Department of Transportation

Research and Innovative Technology Administration, Arlington, Virgina, November

2010

5. “United States Frequency Allocations: The Radio Spectrum,” U.S. Department of

Commerce, National Telecommunication and Allocations Information

Administration, Office of Spectrum Management, October 2003.

6. H. Brueckmann, "Theory and performance of vehicular center-fed whip antenna,"

IRE Trans. on Vehicular Communications , vol. 9, no. 3, pp. 10-20, Dec. 1960.

7. H. Brueckmann, "Improved wide-band VHF whip antenna," IEEE Trans. on

Vehicular Communications , vol. 15, no. 2, pp. 15-32, Oct. 1966.

8. Z. Novacek, "Radiation of a whip antenna on the car body," Radioelectronika, 17th

Int. Conf. , Brno, June 2007, pp. 1-4.

9. H. Jairam, "Novel blade antenna using microstrip patch elements," in Microwave

Conf., 19th European , London, UK, 1989, pp. 1118-1122.

10. A. Walbeoff and R. Langley, "Multiband PCB antenna," Proceedings of IEE

Microwaves, Antennas and Propag. , Centre, Sittingbourne, UK, Dec. 2005, pp. 471-

475.

11. H. Lindenmeier, J. Hopf, and L. Reiter, "Active AM-FM windshield antenna with

equivalent performance to the whip now as standard equipment in car production,"

IEEE AP-S Int. S. , Munich, Germany, June 1985, pp. 621-624.

12. Mu rakami, et. al., “Glass Antenna for Automobile,” United States Patent No.

5,334,988, August 1994.

13. A. A. Heidari, M. Heyrani, and M. Nakhkash, “A Dual -Band Circularly Polarized

Stub Loaded Microstrip Patch Antenna for GPS Applications,” Progress in

Electromagnetics Research, PIER 92, pp 195-208, 2009.

14. F. Mariottini, M. Albani, E. Toniolo, D. Amatori, and S. Maci, "Design of a compact

GPS and SDARS integrated antenna for automotive applications," Antennas and

Wireless Propag. Lett., IEEE , vol. 9, pp. 405-408, May 2010.

15. Yegin, et. al., “Integrated GPS and SDARS Antenna,” United States Patent No.

7,253,770, August 2007.

16. Y. Kim, C. Song, I. Koo, H. Choi, S. Lee, "Design of a double-looped monopole

array antenna for a DSRC system raodside base station," Microwave and Optical

Techno. Lett. , vol. 37, no. 1, pp. 74-77, April 2003.

17. S. Choi, H. Lee, and K. Kwak, “Circularly Polarized H -Shaped Microstrip-Array

Antenna with a T-Slot for DSRC System Roadside Equipment,” Microwave and

Optical Technology Letters, vol. 51, no. 6, pp. 1545-1548, June 2009.

18. G. Z. Rafi, et. al., “Low -Profile Integrated Microstrip Antenna for GPS-DSRC

Application,” IEEE Antennas and Wireless Propagation Letters, Vol. 8, pp. 44 – 48,

2009.

19. K. Wei and Z. Zhang, “Design of a Coplanar Integrated Mi crostrip Antenna for

GPS/ITS Applications,” IEEE Antennas and Wireless Propagation Letters, vol. 10,

pp. 458-461, 2011

20. S. L. Ma, and J.S. Row, “Design of Single -Feed Dual-Frequency Patch Antenna for

GPS and WLAN Applications,” IEEE Transactions on Antennas and Propagation,

vol. 59, no. 9, September 2011.

21. J. S. Belrose, “Fessenden and Marconi: Their Differing Technologies and

Transatlantic Experiments During the First Decade of this Century,” International

Conference on 100 Years of Radio , pp. 86 – 99, September 1995.

22. C. Balanis, Antenna Theory: Analysis and Design, 3 rd ed, Wiley, Hoboken, New

Jersey, 2005.

23. S. Uda, "High angle radiation of short electric waves". Proceedings of the IRE , vol.

15, pp. 377 –385, May 1927.

24. S. Uda and Y. Mushiake, “Yagi-Uda Antenna ”, Research Institute of Electrical

Communication, Tohoku University, Japan, 1954.

25. G. Whyte, “Antennas for Wireless Sensor Network Applications,” Ph. D. Thesis,

University of Glasgow, Glasgow, United Kingdom, 2008.

26. H. T. Hui, K. Y. Chan, and E. K. N. Yung, “The Low Profile Hemispherical Helical

Antenna with Circular Polarization Radiation Pattern over a Wide Angular Range ,”

IEEE Trans. on Ant. and Prop., Vol. 51, No. 6, pp. 1415 – 1418, June 2003.

27. H. Nakano, Y. Okabe, H. Mimaki, and J. Yamauchi , “A Monofilar Spiral Antenna

Excited through a Helical Wire,” IEEE Trans. on Ant. and Prop., Vol. 51, No. 3, pp.

661 – 664, March 2003.

28. T. Buck, et. al., “Tire Pressure Sensor Module and Method for its Manufacture,”

United States Patent No. 20090261962, October 2009.

29. S. Best, “The Radiation Properties of Electrically Small Folded Spherical Helix

Antennas,” IEEE Trans. on Ant. and Prop., Vol. 52, No. 4, pp. 953 – 960, April 2004

30. G. DeJean, “A New High -Gain Microstrip Yagi Array Antenna with a High Front-to-

Back (F/B) Ratio for WLAN and Millimeter-Wave Applications,” IEEE Trans. on

Ant. and Prop., Vol. 55, No. 2, pp. 298 – 304, February 2007.

31. D. Nascimento, R. Schildberg, and J. da S. Lacava, "Low-cost Yagi-Uda monopole

array," IEEE AP-S Int. Symp. , San Diega, CA, July 2008, pp. 1-4.

32. D. Pozar, Microwave Engineering, 3rd ed, Wiley, NewYork, 2005.

33. B. Wadell, Transmission Line Design Handbook, Artech House, Massachusetts,

1991.