Phase-array beamforming fundamental to innovations in 5G RAN

Telecoms.com periodically invites expert third parties to share their views on the industry’s most pressing issues. In this piece Mihai Banu CTO of Blue Danube Systems, examines some of the cutting edge technologies present in 5G radios.

New applications such as remote offices, immersive virtual reality, autonomous vehicles, and telemedicine increasingly fuel the demand for ubiquitous and reliable wireless connectivity. Today, the telecom industry is rapidly expanding the deployment of 5G technologies to meet these demands, in conjunction with the sheer volume of content that now travels over every mobile network.

But for 5G to achieve its true potential, operators need to assess how they maximize the efficiency of spectrum frequencies.

Technology drivers in RAN for 5G and beyond

Historically, the air interface provided by the Radio Access Network (RAN) limits the performance of a wireless cellular network. In principle, 5G RAN is significantly more capable than its predecessors due to the introduction of wider channels in the sub-6 GHz bands, a significant increase in sub-6 GHz spectrum efficiency, the use of mmWave spectrum and expanded cloud orchestration capabilities.

In the traditional Frequency Range 1, or FR1, bands (hundreds of MHz to 6 GHz), channels as wide as 100MHz are now allowed in 5G; a fivefold increase in channel bandwidth compared to 4G. A wider channel supports a higher rate of information flow, producing higher download/upload speeds in the cellular network.

5G also adds FR2 to the traditional frequency range, which spans much higher frequencies than FR1 (e.g., 15-70GHz). MmWave spectrum supports even wider channel bandwidths than FR1 – from hundreds of MHz to GHz – increasing the information flow rate proportionally. However, mmWave bands suffer from high propagation loss and even total link loss in non-line-of-sight transmissions. The transmission range in FR2 is so much smaller than in FR1 that using the conventional RAN design, as in FR1, is not suitable in practice.

Beamforming and phase coherency

Phase-array beamforming can mitigate the range difficulty in mmWave systems by deploying numerous active antenna elements configured in a sizeable dense array to transmit and receive wireless signals coherently (i.e., in an exact phase and magnitude mutual relationship). Controlling the phase and magnitude of the signals at every antenna element creates constructive and destructive electromagnetic interference patterns over the air, generating physical, 3D beams that act like spotlights – a technique used successfully in radars and space exploration.

The net result of using physical beams in 5G mmWave communications is that the significant number of active antennas used multiplies signal strengths, extending the range accordingly. In addition, user separation occurs naturally, as beams illuminate only narrow solid angles. Phased-array beamforming, therefore, achieves spatial multiplexing (transmission of multiple streams of data over the same bandwidth) easily, a powerful way to increase system capacity.

The key to obtaining these beneficial effects in practice is to make sure all active antenna elements operate synchronously; calibrated to high precision under all conditions. But this is a challenging design specification, especially when looking to manage costs.

RF coherency in Massive MIMO systems

I mentioned earlier that 5G significantly increases the FR1 spectrum efficiency (number of bits per Hz). Like mmWave systems, a general approach for achieving this goal is with Massive MIMO (MaMIMO) active antenna arrays, which have 16, 32, or 64 active antenna elements rather than 2, 4, or 8 as in 4G.

There are two general use cases for MaMIMO systems. The primary intention for the first use case is for time-division multiplexing (TDD) systems based on channel sounding and signal processing to obtain the beamforming effects of phased arrays. Channel sounding consists of transmitting overhead pilot signals to measure the channel characteristics, including the radio chains (baseband-to-baseband estimations). Then, using these channel measurements, an appropriate computation creates constructive and destructive interference patterns, just like phased arrays.

Since channel sounding and the computations are done for receive and transmit paths respectively, this method is appropriate for TDD where the two paths are identical (channel reciprocity). Typically, signal boosting and range extension is achieved consistently in TDD systems, but spatial multiplexing is more challenging due to practical errors in channel estimation and hardware impairments. However, frequency-division multiplexing (FDD) systems using this method have shown inferior performance to date due to a lack of channel reciprocity.

The second use case of MaMIMO systems is based on phased-array physical beams and is appropriate for both TDD and FDD. Using physical beams naturally increases the signal strength and range and allows for easy spatial multiplexing. However, the quality of its implementation determines MaMIMO’s performance. Roughly synchronizing and calibrating the array only achieves moderate range extension and little capacity increase. More significant range extension and capacity increases are only possible when the array elements are precisely synchronized and calibrated, such as within a few degrees in phase error and a fraction of dB in magnitude error.

We call this level of precision “RF Coherency” because it produces results practically indistinguishable from ideal phased arrays (zero phase/magnitude errors).

Performance improvements at a lower cost

Recent field trials have shown that it is possible to achieve RF Coherency in 4G and 5G active arrays at a lower cost. Achieving RF Coherency is possible through new RF synchronization and calibration methodologies, implemented with low-complexity, custom mixed-signal integrated circuits, printed circuit board connectivity methods and software/firmware methods.

While this RF Coherency technology applies to all MaMIMO systems (16-64 Tx/Rx), it is compatible with low-cost MaMIMO arrays with reduced radio chains. Here, each radio chain connects to the entire active aperture in contrast with conventional MaMIMO arrays, where each radio chain only connects to a small portion of the active aperture.

Achieving spectral efficiency through advanced AI/ML techniques

Cloud orchestration in traditional RAN, also known as Self Optimized Network (SON) technology, has been mostly limited to simple configuration updates and occasional RF coverage redistribution with Remote Electrical Tilt antennas. The introduction of MaMIMO in 5G and RF Coherency provide an opportunity to enhance Cloud orchestration significantly and deliver tangible network performance improvements.

For example, cloud-based on closed-loop artificial intelligence and machine learning techniques can dynamically and automatically control the precise shape and placement of the 3D physical beams in the second MaMIMO use case I illustrated. The net result is a drastic reduction in cell-to-cell interference and optimum RF energy match to user traffic demands, resulting in a significant capacity increase and better user experience.

Interfaces specified in Open RAN enable the “Super-SON” capability described above. By combining the benefits of Open RAN and RF Coherency technologies, carriers can adopt a highly advanced 5G RAN solution.

Beamforming improvements and Open RAN adoption synergies

Improved spectral efficiency and beam agility using phase coherency and massive MIMO Open RAN technologies create new synergies and opportunities in the global telecoms market by addressing use cases not previously possible. These combined technologies enable various deployment models in very dense and low economy markets like India and LATAM countries. On the other hand, it also has applications for first-world governments and the defense industry, where there is a requirement for higher performance and precision applications.

Beyond 5G, these technologies will be necessary to achieve the 6G key performance goals for enabling applications using ultra-dense deployment topologies and even more challenging spectrum situations.

Dr. Banu is a founder of Blue Danube Systems, which was recently acquired by NEC Corporation. He has over 30 years experience in circuits and systems R&D, with emphasis on analog, radio frequency and mixed-signal integrated circuits. Dr. Banu is the author of more than 30 technical papers, several book chapters and many U.S. and international patents. He received his bachelor’s, master’s, and Ph.D. degrees in electrical engineering from Columbia University and he is an IEEE Fellow.