Disruptive Beamforming Trends Improving Millimeter-Wave 5G
5G is now a reality and the first stage of its infrastructure (sub-6 GHz) is already deployed in major cities around the world. The high data rate demand for 5G mobile users is shown to be fulfilled using the famous Multiple-Input multiple-Output (MIMO) technology . The next deployment stage of 5G is expected to utilize the millimeter-wave (mmWave) frequency spectrum, and the forthcoming base station antennas will operate at frequency bands centered at 28 GHz and 39 GHz. At these high frequencies, a steerable RF beam can reliably serve a communication device in a much better way compared to an inefficient isotropic RF radiator and this is possible by performing beamforming at the base station end, illustrated in Fig. 1. Beamforming is a technique by which a radiator is made to transmit radio signals in a particular direction. A communication device that performs this function is called a beamformer. The most common and simplest type of a beamformer is an array of half-wavelength spaced antennas connected to a single radio frequency (RF) source via a network of power dividers. Such a beamformer is referred to as a corporate-feed array. More sophisticated beamformers involve a bank of phase shifters connected to each antenna element to add beam steering capability to a simple corporate-feed array. Advanced beamformers involve digitally controlled phase shifters, lens structures, intelligent and meta-surfaces, etc., which enhances the beamformer performance.
Fig. 1. mmWave beamformer serving mobile terminals in mmWave 5G network.
Disruptive mmWave Beamforming Technologies:
Designing 5G-ready beamformer hardware at mmWave is challenging due to three major reasons: 1. Huge losses faced by the electromagnetic waves while propagating through the free space, hence highly directive radiation is desirable. 2. The required network of phase-shifters and power dividers to add steering capabilities is lossy and expensive. 3. The theoretical principles of MIMO require each antenna to be connected separately to the baseband processing unit, making the overall system prohibitively expensive, especially when it comes to implementing a 64 or 128 element mmWave massive MIMO system.
In response to these challenges, disruptive technological trends have emerged that are likely to change the way we look at the mmWave beamforming hardware. One such example is the use of a multi-stage lens-based beamformer, in which the requirement of the complex phase shifter and power networks is avoided. As a result, a large number of antennas can be fed using a smaller number of radio frequency chains (power amplifier, mixer, and filter). This way, beamforming gain is achievable, thanks to a large number of antennas, while the cost of the system is kept minimal since the phase-shifting required for beamforming is done in low-cost lens structures. A simple example of such a system is shown below, in which a 15-element antenna array is shown to be capable of generating nine independent radio beams . The system is designed to operate on 28 GHz and is in line with 3GPP standards for 5G. This system is scalable to 64 or even 128 antenna elements, and still, low cost because the beamforming is possible without the requirement of complex and costly phase-shifting networks.
Fig. 2. A 28 GHz two-stage Rotman lens-based beamformer.
A second example is related to successful channel sounding in mmWave 5G bands. The classical radio channel sounder hardware that works well at sub-6 GHz bands of 5G is not efficient enough to support mmWave channels. A new technique of sounding requires much simpler beamforming hardware than the conventional fully connected antenna array and can deliver fast and accurate direction-of-arrival estimations in the mmWave bands. This technique requires only a metallic cavity with sub-wavelength holes on one side and a scatterer placed inside the cavity . An example structure is shown Fig. 3. The cavity uses a frequency-diverse computational approach to do the direction-of-arrival estimation, which requires a single radio frequency chain, hence a low-cost solution again.
Fig. 3. Cavity-backed frequency-diverse antenna for mmWave direction-of-arrival estimation.
A third example is related to mmWave 5G field trials. Although it is always better to rely on channel measurements and field trials to test the practical limits of the mmWave 5G before commercial deployment, rigorous field trials are often not possible and are too expensive to execute. Because of this limitation, the investigation of novel approaches within a network is not possible. In the past, the network planning sector and researchers often relied on a theoretical model to predict network performance. A single antenna used for the network calculations was often considered as an ideal omnidirectional radiator. This approximation was valid because of the simplicity of the system at sub-6 GHz 5G bands.
For mmWave 5G wireless, the assumption of an antenna as an ideal radiator can easily lead to the overestimation of the network performance. The least we can do is to integrate the practically measured 3D beamformer radiation patterns with the theoretical channel models. This approach is even more critical for dense urban environments, where connectivity and reliability of the entire network depend primarily upon the radiation performance of high directivity beamformers. This new technique can reliably estimate the practical mmWave massive MIMO performance by including the measured near-field and far-field 3D radiation patterns into the network calculations that are measured in an anechoic environment like the one shown below.
Fig. 4. mmWave anechoic chamber facility at Queen’s University Belfast.
A fourth example is related to a very large mmWave array hardware. Beamformers at mmWave 5G can operate at full capacity when they have a very large number of radiating antennas. Each antenna is responsible to transmit a fraction of the total available radiated power, which means that each antenna must have a direct or indirect connection to the radio power source. This leads to cumbersome hardware at mmWave frequencies, where technology is not advanced enough to withstand high loss between the radio source and the antennas.
Using sparse antenna arrays is an alternative approach where the total radiated power from the access point is the same, while the number of radiating antennas is less than in a conventional antenna array, in which adjacent antenna spacing must be no larger than λ/2 to avoid grating lobes. Surprisingly, the direction of radiation (main lobe and side lobes) using a sparse antenna array can match perfectly that of a conventional antenna array using the Compressive Sensing  technique. The randomness of antenna locations in a sparse array avoids the introduction of grating lobes while allowing adjacent antenna spacing to be greater than λ/2. This means that a larger array size can be implemented using a relatively small number of antennas.
Fig. 5. A 28 GHz sparse patch antenna array beamformer.
The radio infrastructure required to support mmWave 5G is not ready yet, however, the disruptive technologies are pushing the limits of engineering to make it a reality by 2025. The fastest version of 5G is in fact the mmWave 5G and we are looking forward to the benefits of its ubiquitous ultra-high-speed (up to 10 Gigabits per second) and low latency (down to 0.2 milliseconds).
1. E. Bjornson, L. Van der Perre, S. Buzzi and E. G. Larsson, "Massive MIMO in Sub-6 GHz and mmWave: Physical, Practical, and Use-Case Differences," in IEEE Wireless Communications, vol. 26, no. 2, pp. 100-108, April 2019.
2. M. A. Babar Abbasi, V. F. Fusco and M. Matthaiou, "Millimetre Wave Hybrid Beamforming with Rotman Lens: Performance with Hardware Imperfections," 2019 16th International Symposium on Wireless Communication Systems (ISWCS), Oulu, Finland, 2019, pp. 203-207.
3. Yurduseven, Okan, et al. "Frequency-Diverse Computational Direction of Arrival Estimation Technique." Nature, Scientific reports 9.1 (2019): 1-12.
4. M. A. B. Abbasi, V. Fusco and D. E. Zelenchuk, "Compressive Sensing Multiplicative Antenna Array," in IEEE Transactions on Antennas and Propagation, vol. 66, no. 11, pp. 5918-5925, Nov. 2018.
M. Ali Babar Abbasi is currently Lecturer with the Centre of Wireless Innovation (CWI), Queen’s University Belfast (QUB), Belfast, U.K. He has authored or co-authored more than 50 journal and conference papers and contributed to 4 book chapters. Dr. Abbasi was the grand prize winner at the Mobile World Scholar Challenge at the Mobile World Congress in 2019 (MWC19), Barcelona, Spain. LinkedIn, Researchgate
Vincent F. Fusco (FIEEE, FREng) research focus on advanced microwave through millimeter wave wireless. Personal Chair of High-Frequency Electronic Engineering with QUB. He has authored over 500 scientific papers in major journals and referred international conferences and two textbooks. He holds patents related to self-tracking antennas and has contributed invited papers and book chapters. His current research interests include physical layer secure active antenna techniques. In 2012, he was awarded the IET Senior Achievement Award, the Mountbatten Medal. ResearchGate.