Introduction to Millimeter Wave Wireless Communications
1.1 The Frontier: Millimeter Wave Wireless
Emerging millimeter wave (mmWave) wireless communication systems represent more than a century of evolution in modern communications. Since the early 1900s, when Guglielmo Marconi developed and commercialized the first wireless telegraph communication systems, the wireless industry has expanded from point-to-point technologies, to radio broadcast systems, and finally to wireless networks. As the technology has advanced, wireless communication has become pervasive in our world. Modern society finds itself immersed in wireless networking, as most of us routinely use cellular networks, wireless local area networks, and personal area networks, all which have been developed extensively over the past twenty years. The remarkable popularity of these technologies causes device makers, infrastructure developers, and manufacturers to continually seek greater radio spectrum for more advanced product offerings.
Wireless communication is a transformative medium that allows our work, education, and entertainment to be transported without any physical connection. The capabilities of wireless communications continue to drive human productivity and innovation in many areas. Communication at mmWave operating frequencies represents the most recent game-changing development for wireless systems. Interest in mmWave is in its infancy and will be driven by consumers who continue to desire higher data rates for the consumption of media while demanding lower delays and constant connectivity on wireless devices. At mmWaves, available spectrum is unparalleled compared to cellular and wireless local area network (WLAN) microwave systems that operate at frequencies below 10 GHz. In particular, the unlicensed spectrum at 60 GHz offers 10× to 100× more spectrum than is available for conventional unlicensed wireless local area networks in the Industrial, Scientific, and Medical (ISM) bands (e.g., at 900 MHz, 2.4 GHz, 5 GHz) or for users of WiFi and 4G (or older) cellular systems that operate at carrier frequencies below 6 GHz. To reinforce this perspective, Fig. 1.1 shows the magnitude of spectrum resources at 28 GHz (Local Multipoint Distribution Service [LMDS]) and 60 GHz in comparison to other modern wireless systems. Over 20 GHz of spectrum is waiting to be used for cellular or WLAN traffic in the 28, 38, and 72 GHz bands alone, and hundreds of gigahertz more spectrum could be used at frequencies above 100 GHz. This is a staggering amount of available new spectrum, especially when one considers that all of the world’s cellphones currently operate in less than 1 GHz of allocated spectrum. More spectrum makes it possible to achieve higher data rates for comparable modulation techniques while also providing more resources to be shared among multiple users.
Figure 1.1 Areas of the squares illustrate the available licensed and unlicensed spectrum bandwidths in popular UHF, microwave, 28 GHz LMDS, and 60 GHz mmWave bands in the USA. Other countries around the world have similar spectrum allocations [from [Rap02]].
Research in mmWave has a rich and exciting history. According to [Mil],
- In 1895, Jagadish Chandra Bose first demonstrated in Presidency College, Calcutta, India, transmission and reception of electromagnetic waves at 60 GHz, over 23 meters distance, through two intervening walls by remotely ringing a bell and detonating some gunpowder. For his communication system, Bose pioneered the development of entire millimeter-wave components like: spark transmitter, coherer, dielectric lens, polarizer, horn antenna and cylindrical diffraction grating. This is the first millimeter wave communication system in the world, developed more than 100 years ago.
A pioneering Russian physicist, Pyotr N. Lebedew, also studied transmission and propagation of 4 to 6 mm wavelength radio waves in 1895 [Leb95].
Today’s radio spectrum has become congested due to the widespread use of smart-phones and tablets. Fig. 1.1 shows the relative bandwidth allocations of different spectrum bands in the USA, and Fig. 1.2 shows the spectrum allocations from 30 kHz to 300 GHz according to the Federal Communications Commission (FCC). Note that although Figs. 1.1 and 1.2 represent a particular country (i.e., the USA), other countries around the world have remarkably similar spectrum allocations stemming from the global allocation of spectrum by the World Radiocommunication Conference (WRC) under the auspices of the International Telecommunication Union (ITU). Today’s cellular and personal communication systems (PCS) mostly operate in the UHF ranges from 300 MHz to 3 GHz, and today’s global unlicensed WLAN and wireless personal area network (WPAN) products use the Unlicensed National Information Infrastructure (U-NII) bands of 900 MHz, 2.4 MHz and 5.8 MHz in the low microwave bands. The wireless spectrum right now is already allocated for many different uses and very congested at frequencies below 3 GHz (e.g., UHF and below). AM Radio broadcasting, international shortwave broadcasting, military and ship-to-shore communications, and amateur (ham) radio are just some of the services that use the lower end of the spectrum, from the hundreds of kilohertz to the tens of megahertz (e.g., medium-wave and shortwave bands). Television broadcasting is done from the tens of megahertz to the hundreds of megahertz (e.g., VHF and UHF bands). Current cellphones and wireless devices such as tablets and laptops works at carrier frequencies between 700 MHz and 6 GHz, with channel bandwidths of 5 to 100 MHz. The mmWave spectrum, ranging between 30 and 300 GHz, is occupied by military, radar, and backhaul, but has much lower utilization. In fact, most countries have not even begun to regulate or allocate the spectrum above 100 GHz, as wireless technology at these frequencies has not been commercially viable at reasonable cost points. This is all about to change. Given the large amount of spectrum available, mmWave presents a new opportunity for future mobile communications to use channel bandwidths of 1 GHz or more. Spectrum at 28 GHz, 38 GHz, and 70-80 GHz looks especially promising for next-generation cellular systems. It is amazing to note from Fig. 1.2 that the unlicensed band at 60 GHz contains more spectrum than has been used by every satellite, cellular, WiFi, AM Radio, FM Radio, and television station in the world! This illustrates the massive bandwidths available at mmWave frequencies.
Figure 1.2 Wireless spectrum used by commercial systems in the USA. Each row represents a decade in frequency. For example, today’s 3G and 4G cellular and WiFi carrier frequencies are mostly in between 300 MHz and 3000 MHz, located on the fifth row. Other countries around the world have similar spectrum allocations. Note how the bandwidth of all modern wireless systems (through the first 6 rows) easily fits into the unlicensed 60 GHz band on the bottom row [from [Rap12b] U.S. Dept. of Commerce, NTIA Office of Spectrum Management]. See page C1 (immediately following page 8) for a color version of this figure.
MmWave wireless communication is an enabling technology that has myriad applications to existing and emerging wireless networking deployments. As of the writing of this book, mmWave based on the 60 GHz unlicensed band is seeing active commercial deployment in consumer devices through IEEE 802.11ad [IEE12]. The cellular industry is just beginning to realize the potential of much greater bandwidths for mobile users in the mmWave bands [Gro13][RSM+13]. Many of the design examples in this book draw from the experience in 60 GHz systems and the authors’ early works on mmWave cellular and peer-to-peer studies for the 28 GHz, 38 GHz, 60 GHz, and 72 GHz bands. But 60 GHz WLAN, WPAN, backhaul, and mmWave cellular are only the beginning — these are early versions of the next generation of mmWave and terahertz systems that will support even higher bandwidths and further advances in connectivity.
New 60 GHz wireless products are exciting, not only because of their ability to satisfy consumer demand for high-speed wireless access, but also because 60 GHz products may be deployed worldwide, thanks to harmonious global spectrum regulations. Harmonized global spectrum allocations allow manufacturers to develop worldwide markets, as demonstrated through the widespread adoption and commercial success of IEEE 802.11b WLANs in 1999, and more recent innovations such as IEEE 802.11a, IEEE 802.11g, IEEE 802.11n, and IEEE 802.11ac WLANs that all operate in the same globally allocated spectrum. WLAN succeeded because there was universal international agreement for the use of the 2.4 GHz ISM and 5 GHz Unlicensed National Information Infrastructure bands, which allowed major manufacturers to devote significant resources to create products that could be sold and used globally. Without international spectral agreements, new wireless technologies will founder for lack of a global market. This was demonstrated by early incarnations of Ultra-wide band (UWB) at the turn of the century, whose initial hype dramatically waned in the face of nonuniform worldwide spectral interference regulations. Fortunately, the governments of the USA, Europe, Korea, Japan, and Australia have largely followed the recommendations of the ITU, which designate frequencies between 57 and 66 GHz for unlicensed communications applications [ITU]. In the USA, the FCC has designated bands from 57 to 64 GHz for unlicensed use [Fed06]. In Europe, the European CEPT has allocated bands from 59 to 66 GHz for some form of mobile application [Tan06]. Korea and Japan have designated bands from 57 to 66 GHz and 59 to 66 GHz, respectively [DMRH10]. Australia has dedicated a smaller band from 59.3 to 62.9 GHz. Consequently, there is roughly 7 GHz of spectrum available worldwide for 60 GHz devices.
At the time of this writing, the cellular industry is just beginning to explore similar spectrum harmonization for the use of mobile cellular networks in frequency bands that are in the mmWave spectrum.1 Dubbed “Beyond 4G” or “5G” by the industry, new cellular network concepts that use orders of magnitude more channel bandwidth, for simultaneous mobility coverage as well as wireless backhaul, are just now being introduced to governments and the ITU to create new global spectrum bands at carrier frequencies that are at least an order of magnitude greater than today’s fourth-generation (4G) Long Term Evolution (LTE) and WiMax mobile networks. Thus, just as the WLAN unlicensed products have moved from the carrier frequencies of 1 to 5 GHz in their early generations, now to 60 GHz, the 1 trillion USD cellular industry is about to follow this trend: moving to mmWave frequency bands where massive data rates and new capabilities will be supported by an immense increase in spectrum.
Unlicensed spectrum at 60 GHz is readily available throughout the world, although this was not always the case. The FCC initiated the first major regulation of 60 GHz spectrum for commercial consumers through an unlicensed use proposal in 1995 [Mar10a], yet the same idea was considered a decade earlier by England’s Office of Communications (OfCom) [RMGJ11]. At that time, the FCC considered the mmWave band to be “desert property” due to its perceived unfavorable propagation characteristics and lack of low-cost commercial circuitry. However, the allocation of new spectrum has ignited and will continue to ignite the inventiveness and creativity of engineers to create new consumer products at higher frequencies and greater data rates. This perception of poor propagation due to low distance coverage is heavily influenced by the O2 absorption effect where a 60 GHz carrier wave interacts strongly with atmospheric oxygen during propagation, as illustrated in Fig. 1.3 [RMGJ11][Wel09]. This effect is compounded by other perceived unfavorable qualities of mmWave communication links: increased free space path loss, decreased signal penetration through obstacles, directional communication due to high-gain antenna requirements, and substantial intersymbol interference (ISI, i.e., frequency selectivity) due to many reflective paths over massive operating bandwidths. Furthermore, 60 GHz circuitry and devices have traditionally been very expensive to build, and only in the past few years have circuit solutions become viable in low-cost silicon.
Figure 1.3 Expected atmospheric path loss as a function of frequency under normal atmospheric conditions (101 kPa total air pressure, 22° Celsius air temperature, 10% relative humidity, and 0 g/m3 suspended water droplet concentration) [Lie89]. Note that atmospheric oxygen interacts strongly with electromagnetic waves at 60 GHz. Other carrier frequencies, in dark shading, exhibit strong attenuation peaks due to atmospheric interactions, making them suitable for future short-range applications or “whisper radio” applications where transmissions die out quickly with distance. These bands may service applications similar to 60 GHz with even higher bandwidth, illustrating the future of short-range wireless technologies. It is worth noting, however, that other frequency bands, such as the 20-50 GHz, 70-90 GHz, and 120-160 GHz bands, have very little attenuation, well below 1 dB/km, making them suitable for longer-distance mobile or backhaul communications.
In the early days of 60 GHz wireless communication, many viewed fixed wireless broadband (e.g., fiber backhaul replacement) as the most suitable 60 GHz application, due to requirements for highly directional antennas to achieve acceptable link budgets. Today, however, the propagation characteristics that were once seen as limitations are now either surmountable or seen as advantages. For example, 60 GHz oxygen absorption loss of up to 20 dB/km is almost negligible for networks that operate within 100 meters. The shift away from long-range communications actually benefits close-range communications because it permits aggressive frequency reuse with simultaneously operating networks that do not interfere with each other. Further, the highly directional antennas required for path loss mitigation can actually work to promote security as long as network protocols enable antenna directions to be flexibly steered. Thus, many networks are now finding a home at 60 GHz for communication at distances less than 100 m. Also, the 20 dB/km oxygen attenuation at 60 GHz disappears at other mmWave bands, such as 28, 38, or 72 GHz, making them nearly as good as today’s cellular bands for longer-range outdoor mobile communications. Recent work has found that urban environments provide rich multipath, especially reflected and scattered energy at or above 28 GHz — when smart antennas, beamforming, and spatial processing are used, this rich multipath can be exploited to increase received signal power in non-line of sight (NLOS) propagation environments. Recent results by Samsung show that over 1 Gbps can be carried over mmWave cellular at ranges exceeding 2 km, demonstrating that mmWave bands are useful for cellular networks [Gro13].
Although consumer demand and transformative applications fuel the need for more bandwidth in wireless networks, rapid advancements and price reductions in integrated mmWave (>10 GHz) analog circuits, baseband digital memory, and processors have enabled this progress. Recent developments of integrated mmWave transmitters and receivers with advanced analog and radio frequency (RF) circuitry (see Fig. 1.4) and new phased array and beamforming techniques are also paving the way for the mmWave future (such as the product in Fig. 1.5). Operation at 60 GHz and other mmWave frequencies at reasonable costs is largely the result of a continuation of advancements in complementary metal oxide semiconductor (CMOS) and silicon germanium (SiGe technologies). Signal generation into terahertz frequencies (1 to 430 THz) has been possible since at least the 1960s through photodiodes and other discrete components not amenable to small-scale integration and/or mass production [BS66]. Packaging the analog components needed to generate mmWave RF signals along with the digital hardware necessary to process massive bandwidths, however, has only been possible in the last decade. Moore’s Law, which has accurately predicted that integrated circuit (IC) transistor populations and computations per unit energy will double at regular intervals every two years [NH08, Chapter 1], explains the dramatic advancements that now allow 60 GHz and other mmWave devices to be made inexpensively. Today, transistors made with CMOS and SiGe are fast enough to operate into the range of hundreds of gigahertz [YCP+09], as shown in Fig. 1.6. Further, due to the immense number of transistors required for modern digital circuits (on the order of billions) each transistor is extremely cheap. Inexpensive circuit production processes will make system-on-chip (SoC) mmWave radios — a complete integration of all analog and digital radio components onto a single chip — possible. For mmWave communication, the semiconductor industry is finally ready to produce cost-effective, mass-market products.
Figure 1.4 Block diagram (top) and die photo (bottom) of an integrated circuit with four transmit and receive channels, including the voltage-controlled oscillator, phase-locked loop, and local oscillator distribution network. Beamforming is performed in analog at baseband. Each receiver channel contains a low noise amplifier, inphase/quadrature mixer, and baseband phase rotator. The transmit channel also contains a baseband phase rotator, up-conversion mixers, and power amplifiers. Figure from [TCM+11], courtesy of Prof. Niknejad and Prof. Alon of the Berkeley Wireless Research Center [© IEEE].
Figure 1.5 Third-generation 60 GHz WirelessHD chipset by Silicon Image, including the SiI6320 HRTX Network Processor, SiI6321 HRRX Network Processor, and SiI6310 HRTR RF Transceiver. These chipsets are used in real-time, low-latency applications such as gaming and video, and provide 3.8 Gbps data rates using a steerable 32 element phased array antenna system (courtesy of Silicon Image) [EWA+11] [© IEEE]. See page C2 (immediately preceding page 9) for a color version of this figure.
Figure 1.6 Achievable transit frequency (fT ) of transistors over time for several semiconductor technologies, including silicon CMOS transistors, silicon germanium heterojunction bipolar transistor (SiGe HBT), and certain other III-V high electron mobility transistors (HEMT) and III-V HBTs. Over the last decade CMOS (the current technology of choice for cutting edge digital and analog circuits) has become competitive with III-V technologies for RF and mmWave applications [figure reproduced from data in [RK09]© IEEE].
Wireless personal area networks (WPANs) provided the first mass-market commercial applications of short-range mmWave using the 60 GHz band. The three dominant 60 GHz WPAN specifications are WirelessHD, IEEE 802.11ad (WiGig), and IEEE 802.15.3c. WPANs support connectivity for mobile and peripheral devices; a typical WPAN realization is demonstrated in Fig. 1.7, where products such as those shown in Fig. 1.5 may be used. Currently, the most popular application of WPAN is to provide high-bandwidth connections for cable replacement using the high-definition multimedia interface (HDMI), now proliferating in consumer households. The increasing integration of 60 GHz silicon devices allows implementation on small physical platforms while the massive spectrum allocations at 60 GHz allow media streaming to avoid data compression limitations, which are common at lower frequencies with reduced bandwidth resources. Easing compression requirements is attractive because it reduces signal processing and coding circuity requirements, thereby reducing the digital complexity of a device. This may lead to lower cost and longer battery life in a smaller form factor. Due to major technical and marketing efforts by the Wireless Gigabit Alliance (WiGig), the IEEE 802.11ad standard has been designed to incorporate both WPAN and WLAN capabilities, and WiGig-compliant devices are just starting to ship in laptops, tablets, and smartphones around the world, whereas WirelessHD-compliant devices have been shipping since 2008. The success of today’s USB standard in consumer electronics has demonstrated how harmonious interfaces lead to a proliferation of compatible devices. 60 GHz is poised to fill this role for high-definition multimedia systems, as illustrated in Fig. 1.8.
Figure 1.7 Wireless personal area networking. WPANs often connect mobile devices such as mobile phones and multimedia players to each other as well as desktop computers. Increasing the data-rate beyond current WPANs such as Bluetooth and early UWB was the first driving force for 60 GHz solutions. The IEEE 802.15.3c international standard, the WiGig standard (IEEE 802.11ad), and the earlier WirelessHD standard, released in the 2008–2009 time frame, provide a design for short-range data networks (≈ 10 m). All standards, in their first release, guaranteed to provide (under favorable propagation scenarios) multi-Gbps wireless data transfers to support cable replacement of USB, IEEE 1394, and gigabit Ethernet.
Figure 1.8 Multimedia high-definition (HD) streaming. 60 GHz provides enough spectrum resources to remove HDMI cables without sophisticated joint channel/source coding strategies (e.g., compression), such as in the wireless home digital interface (WHDI) standard that operates at 5 GHz frequencies. Currently, 60 GHz is the only spectrum with sufficient bandwidth to provide a wireless HDMI solution that scales with future HD television technology advancement.
WLANs, which extend the communication range beyond WPAN, also employ mmWave technology in the 60 GHz band. WLANs are used to network computers through a wireless access point, as illustrated in Fig. 1.9, and may connect with other wired networks or to the Internet. WLANs are a popular application of unlicensed spectrum that is being incorporated more broadly into smartphones, tablets, consumer devices, and cars. Currently, most WLAN devices operate under the IEEE 802.11n standard and have the ability to communicate at hundreds of megabits per second. IEEE 802.11n leverages multiple transmit and receive antennas using multiple input multiple output (MIMO) communication methods. These devices carry up to four antennas and operate in the 2.4 GHz or 5.2 GHz unlicensed bands. Until IEEE 802.11n, standard advancements (in terms of data rate capabilities) have been largely linear, that is, a single new standard improves on the previous standard for the next generation of devices. The next generation of WLAN, however, has two standards for gigabit communication: IEEE 802.11ac and IEEE 802.11ad. IEEE 802.11ac is a direct upgrade to IEEE 802.11n through higher-order constellations, more available antennas (up to 8) per device, and up to 4 times more bandwidth at microwave frequencies (5 GHz carrier). IEEE 802.11ad takes a revolutionary approach by exploiting 50 times more bandwidth at mmWave frequencies (60 GHz). It is supported by device manufacturers that recognize the role of mmWave spectrum in the continued bandwidth scaling for next-generation applications. IEEE 802.11ad and mmWave technology will be critical for supporting wireless traffic with speeds competitive not only with gigabit Ethernet, but also 10 gigabit Ethernet and beyond. The largest challenges presented to 60 GHz and mmWave WLAN are the development of power-efficient RF and phased-array antennas and circuitry, and the high attenuation experienced by mmWaves when propagating through certain materials. Many strategies will be employed to overcome these obstacles, including 60 GHz repeaters/relays, adaptive beam steering, and hybrid wired/microwave/mmWave WLAN devices that use copper or fiber cabling or low microwave frequencies for normal operation, and mmWave frequencies when the 60 GHz path loss is favorable. Although the WPAN and WLAN network architectures provide different communication capabilities, several wireless device companies, including Panasonic, Silicon Image, Wilocity, MediaTek, Intel, and Samsung, are aggressively investing in both technologies.
Figure 1.9 Wireless local area networking. WLANs, which typically carry Internet traffic, are a popular application of unlicensed spectrum. WLANs that employ 60 GHz and other mmWave technology provide data rates that are commensurate with gigabit Ethernet. The IEEE 802.11ad and WiGig standards also offer hybrid microwave/mmWave WLAN solutions that use microwave frequencies for normal operation and mmWave frequencies when the 60 GHz path is favorable. Repeaters/relays will be used to provide range and connectivity to additional devices.
MmWave technology also finds applications in cellular systems. One of the earliest applications of mmWave wireless communication was backhaul of gigabit data along a line-of-sight (LOS) path, as illustrated in Fig. 1.10. Transmission ranges on the order of 1 km are possible if very high-gain antennas are deployed. Until recently, however, 60 GHz and mmWave backhaul has largely been viewed as a niche market and has not drawn significant interest. 60 GHz backhaul physical layer (PHY) design traditionally assumed expensive components to provide high reliability and to maximize range, resulting in bulky equipment and reducing the cost advantage over wired backhaul; however, a new application for wireless backhaul is emerging. Cellular systems are increasing in density (resulting in 1 km or less distances between base stations). Concurrently, cellular base stations require higher-capacity backhaul connections to provide mobile high-speed video and to implement advanced multicell cooperation strategies. If wireless backhaul devices are able to leverage recent mmWave hardware cost reductions, they may be able to service this growing need at a lower cost with more infrastructure flexibility. Further, backhaul systems are investigating LOS MIMO strategies to scale throughput into fiber capabilities [SST+09]. As operators continue to move to smaller cell sizes to exploit spatial reuse, the cost per base station will drop as they become more plentiful and more densely distributed in urban areas. Thus, wireless backhaul will be essential for network flexibility, quick deployment, and reduced ongoing operating costs. Consequently, wireless backhaul is likely to reemerge as an important application of 60 GHz and mmWave wireless communications. In fact, we envisage future cellular and WLAN infrastructure to be able to simultaneously handle backhaul, fronthaul, and position location connections using mmWave spectrum.
Figure 1.10 Wireless backhaul and relays may be used to connect multiple cell sites and subscribers together, replacing or augmenting copper or fiber backhaul solutions.
We foresee that mmWave will play a leading role in fifth-generation (5G) cellular networks. In the past generations of cellular technology, various PHY technologies have been successful in achieving ultra-high levels of spectral efficiency (bits/sec/Hz), including orthogonal frequency division multiplexing, multiple antennas, and effi-cient channel coding [GRM+10][STB09][LLL+10][SKM+10][CAG08][GMR+12]. Heterogeneous networks, coordinated multipoint transmission, relays, and the massive deployment of small cells or distributed antennas promise to further increase area spectral effi-ciency (bits/s/Hz/km2) [DMW+11][YHXM09][PPTH09][HPWZ13][CAG08][GMR+12]. The focus on area spectral efficiency is a result of extremely limited bandwidths available in the UHF and microwave frequency bands where cellular systems are deployed, as illustrated in Fig. 1.11. MmWave cellular will change the current operating paradigm using the untapped mmWave spectrum.
Figure 1.11 United States spectrum and bandwidth allocations for 2G, 3G, and 4G LTE-A (long-term evolution advanced). The global spectrum bandwidth allocation for all cellular technologies does not exceed 780 MHz. Currently, allotted spectrum for operators is dissected into disjoint frequency bands, each of which possesses different radio networks with different propagation characteristics and building penetration losses. Each major wireless provider in each country has, at most, approximately 200 MHz of spectrum across all of the different cellular bands available to them [from [RSM+13]© IEEE].
Cellular systems may use mmWave frequencies to augment the currently saturated 700 MHz to 2.6 GHz radio spectrum bands for wireless communications [KP11a]. The combination of cost-effective CMOS technology that can now operate well into the mmWave frequency bands, and high-gain, steerable antennas at the mobile and base station, strengthens the viability of mmWave wireless communications [RSM+13]. MmWave spectrum would allow service providers to offer higher channel bandwidths well beyond the 20 MHz typically available to 4G LTE users. By increasing the RF channel bandwidth for mobile radio channels, the data capacity is greatly increased, while the latency for digital traffic is greatly decreased, thus supporting much better Internet-based access and applications that require minimal latency. Given this significant jump in bandwidth and new capabilities offered by mmWave, the base station-to-device links, as well as backhaul links between base stations, will be able to handle much greater capacity than today’s cellular networks in highly populated areas.
Cellular systems that use mmWave frequencies are likely to be deployed in licensed spectrum at frequencies such as 28 GHz or 38 GHz or at 72 GHz, because licensed spectrum better guarantees the quality of service. The 28 GHz and 38-39 GHz bands are currently available with spectrum allocations of over 1 GHz of bandwidths, and the E-Band above 70 GHz has over 14 GHz available [Gho14]. Originally intended for LMDS use in the late 1990s, the 28 GHz and 38 GHz licenses could be used for mobile cellular as well as backhaul [SA95][RSM+13].
MmWave cellular is a growing topic of research interest [RSM+13]. The use of mmWave for broadband access has been pioneered by Samsung [KP11a][KP11b][PK11] [PKZ10][PLK12], where data rates were reported in the range of 400 Mbps to 2.77 Gbps for a 1 GHz bandwidth at 1 km distance. Nokia has recently demonstrated that 73 GHz could be used to provide peak data rates of over 15 Gbps [Gho14]. Propagation characteristics of promising mmWave bands have been evaluated in [RQT+12], [MBDQ+12], [RSM+13], and [MSR14], and show path loss is slightly larger in NLOS conditions compared with today’s UHF and microwave bands due to the higher carrier frequency. The scattering effects also become important at mmWave frequencies, causing weak signals to become an important source of diversity, and NLOS paths are weaker, making blockage and coverage holes more pronounced. To allow high-quality links, directional beamforming will be needed at both the base station and at the handset where propagation can be improved [GAPR09][RRE14]. Hybrid architectures for beam-forming appear especially attractive as they allow both directional beamforming and more complex forms of precoding while using limited hardware [EAHAS+12a][AELH13]. Applications to picocellular networks are also promising [ALRE13], indicating 15-fold improvements in data rates compared with current 3GPP LTE 4G cellular deployments. Work in [RRE14] shows over 20-fold improvement in end-user data rates over the most advanced 4G LTE networks in New York City. Results in [BAH14] show 12-fold improvements compared with other competing microwave technologies, and results in [ALS+14], [RRE14], and [Gho14] predict 20 times or more capacity improvements using mmWave technologies. As 5G is developed and implemented, we believe the main differences compared to 4G will be the use of much greater spectrum allocations at untapped mmWave frequency bands, highly directional beamforming antennas at both the mobile device and base station, longer battery life, lower outage probability, much higher bit rates in larger portions of the coverage area, cheaper infrastructure costs, and higher aggregate capacity for many simultaneous users in both licensed and unlicensed spectrum, in effect creating a user experience in which massive data-rate cellular and WiFi services are merged.
The architecture of mmWave cellular networks is likely to be much different than in microwave systems, as illustrated in Fig. 1.12. Directional beamforming will result in high gain links between base station and handset, which has the added benefit of reducing outof-cell interference. This means that aggressive spatial reuse may be possible. Backhaul links, for example, may share the same mmWave spectrum, allowing rapid deployment and mesh-like connectivity with cooperation between base stations. MmWave cellular may also make use of microwave frequencies using, for example, the phantom cell concept [KBNI13] where control information is sent on microwave frequencies and data is sent (when possible) on mmWave frequencies.
Figure 1.12 Illustration of a mmWave cellular network. Base stations communicate to users (and interfere with other cell users) via LOS, and NLOS communication, either directly or via heterogeneous infrastructure such as mmWave UWB relays.
A number of universities have research programs in mmWave wireless communication. The University of Surrey, England, has set up a research hub for 5G mobile technology with a goal to expand UK telecommunication research and innovation [Surrey]. New York University (NYU) recently established the NYU WIRELESS research center to create new technologies and fundamental knowledge for future mmWave wireless devices and networks [NYU12]. Aalborg University has an active mmWave research effort. The Wireless Networking and Communications Group (WNCG) at The University of Texas at Austin has a vibrant research program on 5G cellular technologies including mmWave [Wi14]. Aalto University has an active mmWave research effort. The University of Southern California, the University of California at Santa Barbara, the University of California at Berkeley, the California Institute of Technology, the University of Bristol, and the Korea Advanced Institute of Science and Technology (KAIST) are just some of the many universities that have substantial research efforts on mmWave for future wireless networks.
WPANs, WLANs, and cellular communication mark the beginning of mass consumer applications of mmWave technologies, where we evolve to a world where data is transported to and from the cloud and to each other in quantities we cannot fathom today. We believe that mmWave is the “tip of the iceberg” for dramatic new products and changes in our way of life, and will usher in a new generation of engineers and technologists with new capabilities and expertise. This exciting future will bring about revolutionary changes in the way content is distributed, and will completely change the form factor of many electronic devices, motivating the use of larger bandwidths found in the mmWave spectrum for many other types of networks, far beyond 60 GHz [RMGJ11][Rap12a]. For this to happen, however, many challenges must be overcome. Although we predict that future inexpensive UWB wireless cellular and personal area networks will be enabled through a move to mmWave frequencies and continued advancements in highly integrated digital and analog circuitry, we do not predict that all future advancements will be carried on the shoulders of solid-state process engineers, alone. Future wireless engineers will need to understand not only communications engineering and wireless system design principles, but also circuit design, antenna and propagation models, and mmWave electromagnetic theory to successfully codevelop their designs of future wireless solutions.