A few days ago I visited the international business summit in Nizhny Novgorod, at which Megafon and Nokia demonstrated a prototype BS and a 5G standard receiver with a data transfer rate of almost 5 Gbit / s:

In the photo on the right there is a prototype of the base station of the new standard, on the left - the prototype of the subscriber terminal.
')
You do not see that the subscriber terminal is the size of a base station and has an antenna the size of a computer case. By 2018 (namely, the adoption of the 5G standard is expected then), the subscriber part will dry out and, as usual, fit into a small chip in your new phone.
Want to know more? Welcome to the cat: robo-manipulators, a couple of words about how data is transmitted, smart base stations, PTT with video and so on.
For starters, let's get acquainted. Very nice, base station:

Very nice subscriber terminal:

Between them is a five gigabit radio channel:

Well, actually a little less:

Five gigabits is enough to simultaneously distill 8 video streams with a resolution of 7680 Ă— 4320 in real time, plus 3 FullHD video streams and there is still room for a little traffic with minimal delays.

This is achieved by the aggregation of two hundred together with MIMO 8x8. Do not understand? So do I. Therefore, I had to find a person who understands this, and torture him with a dictaphone for half an hour.
Aggregation of two cells means that the data is transmitted not on one frequency, but on two. A cell is a defined carrier supported by a base station. One base station can raise several hundred at once.
Here, for example, the signal spectrum is visible:

On both sides of the central “dip” there are just two carriers, each 100 MHz wide. For example, in 3G networks the carrier width is 5 MHz, and in LTE it is 20 MHz.
MIMO is such a technology for loading into the same band several times more data due to orthogonal codes. In this case, 8x8 is the transmission on the same frequency of 8 different signals for receiving and transmitting.
Articles about OFDM and MIMO are on habr, for example
here and
here .
MIMO is used both to increase speed (as in the case of the current 5G prototype) and to improve communication when the same data is transmitted on different channels. For example, in current 4G (LTE) communication standards there is support for MIMO up to 4x4, although in practice telephones and modems do not use more than 2x2.
The 4.5 GHz band is less penetrating than the 2 GHz band in 3G and 2G, so base stations will have to be installed more often. However, there are advantages in this too - the more stations, the more capacity they have, and the more subscribers can serve, so the situations “a lot of people in one place and nothing works” should be less.

It is important to understand that the current characteristics are not at all the final standard of communication, but only an attempt by Nokia engineers to implement a 5G network according to their own understanding. The final standard is still being developed and approved by
3GPP , and when it is finally fixed in 2018, it will most likely be quite different from the current implementation. However, these experiments will also not be in vain - they will be taken into account when discussing and developing a standard, since real tests are difficult to replace with calculations only.
In addition, the final standard will support the Internet of Things - a variety of small sensors and devices (we leave aside the question of whether it is necessary to connect all these sensors to the cellular network and pay operators for traffic, instead of creating their own network).
So how do we show speed and low latency?
For example, here is the platform on which the ball freely rolls, supported by three manipulators:

They are controlled by a computer that receives an image from a camera mounted above the platform. Seeing the ball, he forms commands for the movement of the manipulators so that the ball rolls right into the center of the platform. How does the task from the field of machine vision relate to 5G networks?
And the whole point is that the video data and commands to the manipulators are transmitted through the radio bridge “base station - subscriber terminal”. Accordingly, there is a certain delay between sending data from the camera and receiving the commands by manipulators, which is caused by the image processing and data transfer speed (more precisely, not even by speed, but by the delay).
If we artificially introduce a delay of about 100ms into this channel, imitating the 4G connection, then we observe the following picture:

Without proper optimization at the time of receiving commands to change the position of the platform, the ball position is no longer relevant - he managed to roll out of the calculated trajectory, so the team that reached the manipulator only makes it worse, forcing the ball to move in the wrong direction. Thus, a delay of 100ms, poorly sensed by a person, already sufficiently affects even such a simple system as a rolling ball on a board.

But it is worth removing the delay introduced by switching the parameters of the data channel to the real parameters of the 5G technology (5ms at full channel load, without loading two times less, the delay includes passing the signal back and forth), as everything magically returns to normal: after a few seconds, the control program stops the ball in the center of the platform.

Virtual reality also applies to 5G networks: the transmission of data on the position of the glasses and the transfer of the picture back to the glasses are also carried out through the 5G network, which demonstrates low delays in video transmission, imperceptible to the human eye.

Of course, in real life, no one would ever think of making virtual reality glasses with transmitting video over a cellular network, this is just an example showing the capabilities of the technology. I put on my glasses and turned my head with pleasure as I flew past Moscow City. I confirm, the delays are not noticeable. The pixels with glasses are noticeable, but there is no delay.
Another booth is dedicated to the so-called “smart base stations” - Mobile Edge Computing technology. This is a technology that allows you to install servers in base stations that are engaged in some tasks “without departing from the cash register”. This ensures low latency (just those 5ms, because everyone understands that in real use, the signal travel time to third-party servers and back will be added to these 5ms), and saving channels from the BS to the provider.
For example, here on one of these servers the motion detector on video works:

You go inside the protected perimeter, it starts beeping:

I haven’t yet figured out where this technology might be useful, but let it be.
The latest exposition is protected communicators with real-time audio and video communication on the Missian Critical Push-To-Tallk solution:

If anyone does not know, PTT is an attempt to make a walkie-talkie out of cell phones. You do not need to dial the number, press the button, and all members of your group can hear you (which is no longer limited by the radio signal strength of the radio). In addition to audio, there is also video:
- Petrovich, look at my wire torn off. What colors to combine?In addition, the advantages are that you can record all the negotiations and display the transmission to the dispatcher on the computer.
On this look interesting exhibits finish, eat cookies:

And we go back to Moscow, to our native 4G.
