Last month I had a post about 5G heavily relying on spectrum beyond 5 GHz and the catch that today's consumer devices using such spectrum can only cover a few meters. In other words, using spectrum in the 30 or even 70 GHz range (called millimiter-waves, or mmW for short) won't work for cellular networks where base stations are several hundreds of meters apart from each other even in dense population areas. Fortunately, not everybody agrees.
After I posted my article I was made me aware of this very interesting IEEE article in which the authors describe their study of how such ultra high frequency ranges could become usable in a cellular environment. Their conclusion, based on experiments in a real environment, is that by using a high number of tiny antennas for beamforming in mobile devices and base stations, it's possible to overcome the high attenuation of the air interface in the 30 and 70 GHz bands and thus significantly increase the transmission range. They predict that the combination of beamforming and using large 1 GHz carriers can increase overall air interface capacity by an order of a magnitude compared to the 20 MHz carriers used for LTE today.
As antennas are small, a space in a smartphone of 1.5 to 1.5 centimeters could hold 16+ tiny antennas which would be enough to achieve the desired beamforming effect. The authors note, however, that with the current approach of treating each signal path separately it is not feasible to process so many inputs and outputs and that new methods have to be found, especially on the smartphone side, to master this new level of complexity while keeping the amount of energy necessary for the processing in check.
Another challenge pointed out by the authors is indoor coverage, because even with beamforming, millimeter waves are still not penetrating walls and other solid obstacles well. In other words, mmW base stations must also be put inside buildings to also go beyond today's data rates there. In many cases it's unlikely that several operators can deploy their mmW equipment inside a single building so the authors note that a new business model might be required where a third party offers mmW access equipment for interconnection to traditional mobile backhaul networks.
While beamforming holds the solution to extending the range of millimeter-wave systems to usable distances one issue that is created by this is how synchronization and broadcast channels that have to be transmitted omnidirectionally can reach devices. In addition, the channel state from and to each device needs to be continuously tracked in order to keep the beams aligned to individual devices. This is likely going to impact a device's power saving abilities as the transmitter can't be off for long periods even if no data has to be transmitted.
The authors list many further points that have to be considered due to the completely different nature of using mmW in combination with beamforming compared to today's LTE systems. It's by far not a drive-by read so bring some time if you want to explore the paper, it's definitely worth it.
Thanks to Guy Daniels for pointing this the paper via his article on the topic here!
I’m currently in Geneva at the World Radiocommunication Conference, where we are considering what frequency bands to study with a view to defining new allocations for 5G type services (IMT) in the future (by the way, it’s 23:30 and there’s no end in sight for the discussion!) Bands in the range 10-50GHz are indeed being looked at so advanced beam forming antennas will be key to future devices. But these allocations won’t happen until the studies are done to show that IMT can coexist with other radio services in the same or adjacent frequencies – that will be 2019 at the earliest.