From mobile telephones to wireless Internet access to networked appliances and peripherals, there is an increasing reliance on wireless communications to provide or enhance functionality for products and services. The wireless communications industry continues to generate new products and applications for consumers and new opportunities for businesses. The enormity of the opportunities presented by the current explosion of wireless applications is accompanied by comparable design and manufacturing challenges.
The world of wireless communications can appear to be a confusing offering of services, products and standards, all competing for dominance in a dynamic marketplace. But regardless of modulation, protocol, bandwidth or frequency, every wireless device requires an antenna for transmission and/or reception. The antenna is often taken for granted, but its performance is critical to the successful operation of any wireless system. Three major areas of activity in antenna research and development have emerged to meet the needs of modern communications systems: size reduction, wideband or multi-band operation, and adaptive pattern control.
As communication devices become smaller due to greater integration of electronics, the antenna becomes a significantly larger part of the overall package volume. This results in a demand for similar reductions in antenna size. However, reducing antenna size without significantly impacting gain and efficiency is a challenging task.
As integration increases, a single antenna is often required to support two or more of the many wireless services across a broad frequency range. Multi-band and wideband antennas are being developed to meet this need.
The third trend is the increased use of antenna arrays and development of new approaches for using arrays to improve system performance. These new compact, multi-band, wideband and array antennas will serve in a wide variety of applications, but require innovative designs.
In addition to the familiar cellular and PCS voice services, there are many other audio, data and video applications. Wireless local loop service may finally become available using the 2.4 GHz ISM or 5.7 GHz NII bands. Wireless services include entertainment as well as voice communications. For example, high fidelity digital audio is available to fixed and mobile users in the continental US via direct broadcast satellite, called Satellite Digital Audio Radio Service (SDARS). Short-range wireless data communications are likely to become ubiquitous as inexpensive transceivers using the Bluetooth standard are embedded in a wide range of devices. Wireless local area networks (WLAN) have existed for several years, but now large wireless data networks are under development. One example is a county-wide buildout in Utah of a packet data network using PC card-based modems and compact antennas.1 Third-generation cellular standards will provide sufficient bandwidth to support video and data as well as voice service. These and other emerging wireless applications will require high performance, low profile antennas to operate in fixed, mobile, handheld and airborne environments.
In this article, the current and future trends in antennas for wireless communications and some of the innovative approaches being used to meet the demand for antennas that have small size, wide bandwidth, high gain and/or adaptive pattern control are examined.
Antenna Requirements and Designs Tools
Several parameters are important to antenna performance, but antenna gain, pattern, polarization, bandwidth, size and cost are the most important.2 The design and development process begins with a complete list of the specific performance parameter goal values for the application, followed by extensive engineering investigations to select a final design.
The response of a receiving antenna to an incident plane wave as a function of the angle of arrival of the wave is referred to as the antenna pattern . The transmitting pattern of an antenna is the same as the receiving pattern. Antenna beamwidth is the angular extent of the main lobe of the pattern in one plane (typically measured at 3 dB below the maximum) and is inversely proportional to gain . Antenna (power) gain is the product of antenna radiation efficiency and directivity, which is the power increase in the pattern peak direction compared to a uniform angular distribution of radiated power. For fixed point-to-point links, both terminals usually have directional antennas with high gain. Point-to-multipoint links, such as for mobile communications, usually use low gain antennas. For example, a sector antenna for cellular communications must provide a broad azimuth beam that covers the 120° angular extent of a sector (the 3 dB beamwidth may be somewhat narrower than 120° in practice because the roll-off of the main beam is gradual). Such an antenna will have a much lower gain than an antenna that is used for a point-to-point microwave link.
Another key antenna characteristic is polarization , which refers to the orientation of the radiated electric field vector at a point in space as a function of time. The most common types of antenna polarization are linear (vertical or horizontal) and circular (right- or left-hand sense). For every antenna there is a polarization state that is orthogonal to the antenna's polarization, and the antenna does not respond to energy in that polarization. This property can be exploited to reuse a frequency on two orthogonal polarizations. In practice, it is not possible to achieve perfect orthogonality, but adaptive interference cancellers can reduce interference between the cross-polarized channels. This type of polarization reuse is only practical on clear line-of-sight links. Circularly polarized (CP) antenna designs are generally more complicated than linearly polarized designs, but can provide multi-path mitigation in line-of-sight channels. If matched CP antennas are used for transmitting and receiving, single-bounce reflections will arrive at the receiver with a polarization nearly orthogonal to the receiving antenna polarization. The received reflected signals will not be strong enough to cancel the direct signal, avoiding so-called multi-path effects.
A third important parameter is bandwidth , the range of frequencies over which an antenna meets a specified performance threshold. Typically, the antenna bandwidth is limited by impedance match, gain, or polarization. A poorly matched antenna reflects a significant amount of power back to the transmitter. This reduces the radiated power and can harm the transmitter if the power level is high. A poor impedance match also causes much of the received signal power to be reflected from the receiver back towards the antenna. Bandwidth can be specified as a percentage of the center frequency, but for very broadband antennas it is often specified as a ratio of high to low frequencies.
For many antenna applications, such as handheld transceivers, small size is extremely important. This leads to performance tradeoffs because both gain and bandwidth diminish as the size of an antenna is reduced. It is also difficult to achieve nearly pure linear or circular polarization with a small antenna. These important design tradeoffs can now be investigated using software design tools.
A new antenna design often begins as a variation of an existing design, a combination of features of two or more antenna types, or an entirely new idea, and an understanding of electromagnetics. The basic design idea must be refined to optimize performance, and both software modeling and measurements are useful for this refinement. Several software packages exist that can be used to perform electromagnetic modeling of antennas using a variety of methods. These include moment-method codes such as Numerical Electromagnetics Code (NEC), MININEC3 and WIRE,4 finite element method (FEM) software, including HFSS,5 and finite-difference time domain (FDTD) codes such as Fidelity.6 Additional programs have been developed to model array performance. These software packages are used to model and evaluate performance of candidate antenna designs. Trends in antenna performance vs. design parameter values can be observed and designs can be refined in an iterative process before the antenna is actually constructed. Measuring the characteristics of the physical antenna, such as impedance and radiation pattern, is a necessary step and can lead to further iterations. Designs that are selected for mass production must be further refined during the pre-production development cycle to support manufacturability.
Modern applications are pushing the limits of antenna performance. Reduced-size antennas are increasingly desirable for both fixed and mobile applications. Small and handheld devices require low profile antennas, and when possible the antennas are embedded in the device. For most applications (ultra wide band (UWB) is an exception) the communications channel bandwidth does not approach the antenna bandwidth, but broadband or multi-band antennas may be required to cover two or more bands in order to support multiple services. In both commercial and military applications it is desirable to minimize the number of antennas while maintaining high performance for all wireless applications, so that a phone, computer, car, plane, or ship does not have to carry several antennas. While high gain antennas are desirable for fixed point-to-point use, mobile and handheld applications demand lower gain omnidirectional antennas that are small but efficient. In some applications multiple antennas are used in an array to provide high gain, diversity, or interference rejection. All these antennas are designed through a combination of understanding of electromagnetic principles and an iterative process of parameter variation, computer modeling, and prototype construction and measurement.
Low Profile and Embedded Antennas
As new radio handsets become smaller and as transceivers are integrated into other devices, more compact antennas are needed. Unobtrusive reduced-size antennas are also desirable for fixed applications such as micro- and pico-cellular base stations and wireless data network hubs. These antennas must provide sufficient gain and bandwidth, with directional or omnidirectional patterns suited to their intended applications. Two major challenges arise in the design of small antennas. First, there is a fundamental relationship between the size, bandwidth and efficiency of an antenna. As antennas are made smaller, either the operating bandwidth or the antenna efficiency must decrease. Second, gain is related to the size of the antenna; that is, small antennas typically provide lower gain than larger antennas. There is a push to develop low profile and embedded antennas throughout the wireless communications industry for a variety of applications. Some illustrations based on several antenna designs are shown in this article.
Many types of low profile, compact antennas with high efficiency are available for use in compact terminals. They include the inverted-L (ILA), inverted-F (IFA), dual inverted-F (DIFA) and planar inverted-F (PIFA) antennas. These antennas have bandwidths from a few percent for the ILA to about 10 percent for the PIFA. Ways to extend the bandwidth of antennas in the PIFA family of antennas without performance sacrifices have been investigated. One such compact antenna that operates near the fundamental size-bandwidth limit is the wideband compact planar inverted F antenna (WC-PIFA). A WC-PIFA for 2.4 GHz is shown in Figure 1 . It provides an omnidirectional pattern and a 2:1 SWR bandwidth of approximately 40 percent of the center frequency or a bandwidth of approximately 1.5:1, as seen in Figure 2 . Based on computer modeling, the WC-PIFA has a nearly omnidirectional pattern with a maximum gain of 2 to 3 dBi across its bandwidth. The WC-PIFA can easily be embedded in a radio handset, PCMCIA wireless modem, or other wireless device. The WC-PIFA is detuned only slightly by the presence of a human hand in close proximity to the antenna.
In addition to the obvious requirement for small antennas on handheld terminals, low profile antenna designs are important for fixed wireless applications. Many localities are becoming more restrictive in zoning requirements for towers and antennas in an effort to reduce visual clutter within their communities. Antenna designs that are lower profile or physically smaller stand a much better chance of blending into the background. Additionally, smaller antennas present much lower wind loading, which means that lighter support structures are required. This reduces installation costs as well as improves the antenna durability in harsh weather conditions.
One example of a low profile, high performance antenna for fixed applications is the stub loaded helix (SLH) antenna.7 The SLH, shown in Figure 3 , is derived from the axial mode helix antenna, and like the axial mode helix, it produces circular polarization and high gain. The SLH uses a unique geometry that produces comparable gain performance to the conventional helix but in a volume approximately one-fourth the size. This significant reduction in size coupled with its high performance makes the SLH antenna very attractive for use in point-to-point links and for point-to-multipoint coverage without causing significant visual impact. The SLH antenna has been commercialized for use in 2.4 GHz WLAN systems.1 Typical gain from an SLH antenna with an axial length of 8" at 2.4 GHz is 10 dBic.
Some small antennas are backed by a ground plane to reflect signals and achieve a directional pattern. Typically, the ground plane is made of an electrical conductor and is located a quarter wavelength from the antenna so that direct and reflected waves add in-phase. This quarter wavelength spacing can add significantly to the antenna size. Artificial magnetic conductors, or electromagnetic band gap structures, have been developed that provide an in-phase reflection when mounted very close to the antenna. This results in an antenna with reduced size, at the expense of another desirable characteristic, bandwidth.
Wideband and Multi-band Antennas
Mobile users often access several services that cover a wide frequency range. Examples of popular mobile communication and position location services are: cellular at 800 MHz, PCS and GSM at 1900 MHz, Global Positioning Satellite (GPS) at 1500 MHz, and the unlicensed 2400 MHz band. These are in addition to the conventional AM/FM broadcast services traditionally found on vehicles. Soon services such as wireless Internet access, intelligent vehicle systems and satellite digital audio broadcast will be available to mobile users.
Traditionally, the addition of a new communications service to a vehicle required adding another antenna, typically some variant of the whip monopole. This usually produces a porcupine look for high tech vehicles for public and commercial service, but this look is unacceptable to the consumer public. Automobile manufacturers are integrating more communications electronics into new models. Cars of the future must have unobtrusive antennas that support these many functions, requiring antennas with multi-band or wideband capabilities. Automobile manufacturers would like to have a suite of low profile antennas in one or a few locations on the vehicle to support multiple wireless services.
Antennas that cover more than one band are already used on newer mobile phones that support conventional cellular and PCS/GSM. Cellular/GPS antennas are available as well.8 This trend toward multi-band capability will continue and accelerate as more services at different frequencies become available.
Multi-band antennas are usually designed to provide a good impedance match and performance over two or more specific, narrow frequency bands. As services and frequency requirements proliferate, wideband antennas may be a more economical solution. Wideband antennas perform consistently across a continuous block of spectrum, providing capabilities for current and future applications that are not limited to specific narrow bands. An additional type of antenna that is emerging is the reconfigurable antenna that can be tuned over a frequency range that is wider than the instantaneous bandwidth of the antenna.
Many wide bandwidth antennas belong to a class known as frequency independent antennas. They have a self-similar structure that varies in some manner, usually logarithmically, such that the active part of the antenna looks electrically identical at each frequency. Typical of these types of antennas are the log-periodic dipole array, the spiral and the sinuous antenna. The spiral and the sinuous antennas, shown in Figure 4 , are capable of providing a good impedance match and relatively consistent gain performance over several octaves or more of bandwidth. The spiral produces circular polarization exclusively, whereas the sinuous can produce vertical, horizontal, or circular polarization at the expense of a more complicated feed structure. These antennas have a dimension of approximately 0.5 
at the lowest operating frequency.
Recently, more compact wide bandwidth antennas have emerged, such as the Foursquare and the Fourpoint antennas.9,10 The Foursquare (Figure 5 ) has a 2:1 SWR bandwidth of 1.8:1 and the Fourpoint (Figure 6 ) has nearly a 3:1 measured bandwidth. The figures do not show details of the Foursquare and Fourpoint antennas; the resemblance to a simple narrowband microstrip patch antenna is superficial. Both Foursquare and Fourpoint antennas provide dual orthogonal linear polarizations and, like the sinuous antenna, can be used with a suitable feed to transmit or receive waves having any desired polarization state. Both of these antennas occupy a square that has sides that are 0.3 to 0.4 wavelength in length, at the lowest operating frequency. A single one of these antennas provides a nearly symmetric unidirectional beam pattern that is constant over a wide bandwidth. The gain of the foursquare antenna backed by a ground plane has been measured over part of the operating bandwidth and is about 7 to 9 dB. For higher gain and narrower beamwidth, multiple elements are used to form an array.
Arrays and "Smart" Antennas
The future of wireless systems will certainly include wider deployment of arrays. Among these arrays are so-called "smart" antennas, which integrate radio intelligence with the antenna. Antenna arrays that have been widely used in military applications for decades are beginning to find commercial applications. Arrays use multiple antennas, or elements, to achieve enhanced performance including high gain. They can also support electrical beam steering to improve transmission and reception, and null steering to reject interfering signals. As with single antennas, arrays are sometimes required to cover a wide bandwidth.
Military ships and aircraft have limited space for onboard antennas. They require arrays that support communications, radar, signal intelligence and navigation across a wide range of frequencies. Similar solutions are needed for civilian vehicles that carry radios for communications, navigation and entertainment. Design of these wideband, multifunctional arrays is challenging. Wideband elements like those discussed in the last section are needed to form the basis of the array. The elements must be arranged carefully to allow the main beam of the array to be scanned over a wide angular range at all frequencies within the array bandwidth. If the element spacing is too large, then undesirable grating lobes (antenna beams in directions other than the desired beam direction) will arise when the array beam is steered. The element spacing must also accommodate elements that are large enough to operate at the low end of the array's bandwidth. Electromagnetic interactions between closely spaced array elements distort the pattern of the individual elements and affect the achievable array patterns as well. This mutual coupling, a frequency dependent effect, is an important consideration.
When a single array configuration does not meet the demands of an application, a reconfigurable array can be used. An early example is the Wullenweber array,11 a circular array developed for direction finding at HF frequencies. The array can use either omnidirectional elements or directional elements that are oriented radially outward. The array typically consists of 30 to 100 evenly spaced elements. A contiguous set of approximately a third of the elements is used to form a beam that is oriented radially outward from the array. A switching network called a goniometer is used to connect the appropriate elements to the radio, and may include some amplitude weighting to control the array pattern. Newer approaches use PIN diode or microelectromechanical system (MEMS) switches to reconfigure arrays or elements within the array for specific frequency bands or operational scenarios.
Array antennas can be used in conjunction with signal processing techniques such as spatial and space-time adaptive processing. In these approaches the array pattern is dynamically controlled to optimize the received signal. In the case of space-time processing, the pattern control is performed jointly with channel equalization to help mitigate interference between symbols that occurs when wideband digital signals are transmitted over multi-path channels. Another advanced technique that uses multiple antennas is space-time coding. By using more than one antenna at the transmitter and at the receiver, a communications system with space-time coding effectively exploits multiple spatial channels to support communications at data rates beyond those predicted by the Shannon limit for a single channel.
Future wireless applications may see antenna arrays integrated into the body panels of vehicles providing the wideband coverage needed for the myriad mobile services being deployed. Since it is difficult to obtain omnidirectional coverage on a vehicle, multiple elements can be placed around the vehicle body to provide the desired coverage. An example of this in Figure 7 shows a scale model of an automobile using antennas placed within the sideview mirror enclosure to provide coverage of the forward hemisphere.
Besides combining the elements to form a single beam, they may be used individually to provide diversity gain in channels that exhibit fading due to multi-path interference. Multiple antennas are used at cellular base stations to provide diversity on both transmit and receive in order to improve coverage and reliability within the cell.
At base stations, wide spacing (10 to 20 wavelengths) between elements is usually required in order to achieve low correlation between signal fading envelopes at different antennas, which is a necessary condition for significant diversity improvements. Recent research has demonstrated that diversity gain can be realized using multiple antennas mounted in very close proximity on a handheld radio. This is possible because of the wide angle spread of multi-path components for a handheld terminal operating in a cluttered environment. Using a combination of spatial and polarization diversity, meaningful improvements in signal fading have been observed.
A smart base station testbed was developed that allows direct comparison of spatial, polarization and angle diversity. The testbed, shown in Figure 8 , operates in the cellular band. In addition to the multi-channel base station receiver and data logger, the testbed includes a transmitter that can be hand-carried or mounted in a vehicle. The testbed was used to measure diversity gain with a transmitter in urban and suburban locations at distances of 665 to 2670 m from the base station antennas. Diversity gains of 4.6 to 10.9 dB at the 1 percent probability level were achieved using selection diversity with spatial, polarization and angle diversity antenna configurations.12
Figure 9 shows top and side views of a handheld array that is part of another testbed. This handheld antenna array testbed (HAAT) was used to measure a diversity gain of 7 to 9 dB at the 1 percent cumulative probability level for spatial, pattern and polarization diversity, using two diversity branches.13 The HAAT was also used to evaluate adaptive beamforming using handheld arrays. Four-element multi-polarized arrays achieved mean signal-to-interference-plus-noise ratios (SINR) of 25 to 50 dB after beamforming in peer-to-peer line of sight scenarios, and 12 to 26 dB in microcellular scenarios.
Figure 10 shows results from an indoor experiment in which data were recorded while a receiver with an array of four half-wave dipole antennas with 0.17 spacing was carried through a hallway. Two co-channel transmitters, one providing a desired signal and the other an interfering signal, operated in rooms off the hallway. From the cumulative distribution function, it can be seen that SINR for the desired signal was improved by 25 to 30 dB (the horizontal distance between the Ch. 1-Ch. 4 curves on the left and the Output After Beamforming curve on the right) using a digital beamformer. Beamforming was performed using a variation of the multi-target constant modulus algorithm.14 The two testbeds described here use narrowband signals. Future research will use wideband signals and spatial-temporal techniques that combine equalization with diversity and beamforming.
Conclusion
New high performance antennas are being developed to satisfy the competing demands of emerging wireless applications for small size, for embedded and low profile applications, and broad bandwidth or multi-band capability to support multiple services. In addition, array antennas can be used to steer beams and nulls as needed to support communications. Arrays are combined with signal processing to implement diversity combining, adaptive beamforming or space-time processing to overcome the effects of multi-path propagation and interference. Multiple antennas can also be used to increase channel capacity with space-time coding. Reconfigurable antennas, feed networks and entire arrays will serve in applications in which a single conventional antenna or array is not suitable. These structures can be adapted dynamically using PIN diode or MEMS switches, and provide great flexibility. Over a century after Marconi's and Popov's first radio experiments, the dynamic field of antenna research and development to support wireless communications remains active, and intensive work to find antennas that approach fundamental performance limits continues.
References
1. Turbowave Inc. (http://www.turbowave.com)
2. W.L. Stutzman and G.A. Thiele, Antenna Theory and Design , Second Edition, John Wiley & Sons, New York, NY 1998.
3. MiniNEC is a product of EM Scientific Inc. (http://www.emsci.com)
4. WIRE is a freeware moment method code written by W.A. Davis. Further information is available at http://filebox.vt.edu/ eng/ee/faculty/WDavis/antenna/.
5. HFSS is a product of Ansoft Corp. (http://www.ansoft.com)
6. Fidelity is a product of Zeland Software Inc. (http://www.zeland.com)
7. U.S. Patent 5,986,621, Stub Loaded Helix Antenna , November 16, 1999.
8. N. Cummings, "Low Profile Integrated GPS and Cellular Antenna," Master's Thesis, Virginia Polytechnic Institute and State University, September 2001.
9. U.S. Patent 5,926,137, Foursquare Antenna Radiating Element , July 20, 1999.
10. U.S. Patent 6,057,802, Trimmed Foursquare Antenna Radiating Element , May 2, 2000.
11. H.D. Kennedy and R.B. Woolsey, "Direction-finding Antennas and Systems," Antenna Engineering Handbook , Third Edition, R.C. Johnson, editor, McGraw-Hill, 1993, Chapter 39.
12. B.K. Kim, W.L. Stutzman and D.G. Sweeney, "Indoor and Outdoor Measurements of Space, Polarization and Angle Diversity for Cellular Base Stations in Urban Environments," IEEE Vehicular Technology Conference, 2000, Vol. 1, pp. 22-29.
13. C.B. Dietrich, K.Dietze, J.R. Nealy and W.L. Stutzman, "Spatial, Polarization and Pattern Diversity for Wireless Handheld Terminals," IEEE Transactions on Antennas and Propagation , Vol. 49, No. 9, September 2001, pp. 1271-1281.
14. B. Agee, "Blind Separation and Capture of Communication Signals Using a Multitarget Constant Modulus Beamformer," IEEE Military Communications Conference, 1989, pp. 340-346.
C.B. Dietrich, Jr. received his BSEE degree from Texas A&M University and his MS and PhD degrees in electrical engineering from Virginia Tech. He is currently a research associate in the Virginia Tech Antenna Group. His research interests include adaptive arrays and diversity systems for reliable wireless communications.
R.M. Barts received his BSEE and MSEE degrees from Virginia Tech. His current research focus is on antennas for new wireless LAN systems. He is the co-recipient of the 1992 Wheeler Award from the IEEE Antennas and Propagation Society, and has been awarded a patent for the stub loaded helix antenna.
W.L. Stutzman received his BS degree in electrical engineering and his AB degree in mathematics from the University of Illinois in 1964, and received his MS and PhD degrees in electrical engineering from Ohio State University in 1965 and 1969, respectively. Dr. Stutzman has been on the electrical engineering faculty of Virginia Polytechnic Institute and State University since 1969, and is currently the Thomas Phillips Professor of Engineering. He is a fellow of the IEEE and served as president of the IEEE Antennas and Propagation Society in 1992.
W.A. Davis received his PhD degree from the University of Illinois. He is strongly involved in numerical methods for electromagnetics as well as antennas, microwave measurements and material characterization. He is the director of the Virginia Tech Antenna Group (VTAG), which is part of the Center for Wireless Telecommunications (CWT), with an emphasis in array technology, low profile, broadband antennas, propagation and measurements.
