Jonathan Newell talks to National Instruments about the cutting edge research taking place to create the next networking generation.
It takes about half a human generation to develop a wireless network generation and with 5G, we’re already on the fifth. Work was started on the 5G standard only three years ago so, based on previous history, we can expect fifth generation deployments to be commonplace by 2025. However, with predictions of up to 20 billion devices being connected on the internet of things (IoT) by 2020, can we wait until the middle of the next decade for a suitable network on which to connect everything?
To find out more about 5G and what research and development work is taking place, I spoke to National Instruments’ Software Defined Radio (SDR) product marketing manager, Sarah Yost. An electrical engineer with a background in microwave and millimetre wave (mmWave) technology for wireless communications, Yost has a deep knowledge of 5G prototyping and SDR.
5G Technology Directions
Since the kick off a few short years ago, there have been three main areas of focus in 5G development, including:
1 Enhanced mobile broadband for commercial operators
2 Massive machine control to cope with the expected explosion in IoT devices
3 Ultra-reliable machine control for such technology as autonomous vehicles
According to Yost, there isn’t a single technology thread under the 5G umbrella that caters for all three of these areas. However, the ones that are receiving the bulk of attention are the use of the mmWave spectrum for obtaining additional bandwidth and “Massive MIMO” for enabling the more effective use of existing bandwidth.
Mobile operators, equipment specialists and academic institutions all over the world are working on these topics and National Instruments is working with telecom operators as well as such academic institutions as New York University (NYU), Lund University and the University of Bristol, which is working with BT.
Such intense research reflect the demanding nature of the standard and the need for swift progress. “So much innovation is happening in the field at the moment because the 5G standard is so demanding,” Yost told me.
NI is working collaboratively with the Universities of Bristol and Lund on massive MIMO field trials, such as the BT project. The idea of Massive MIMO is to gain spectral efficiency improvements though the use of multiple transmission antennas. MIMO (Multiple Input Multiple Output) is nothing new in itself as BT base stations typically have 8 antennas, however Massive MIMO multiplies this significantly.
“As part of the work Bristol is doing with BT, a base station has been equipped with 128 antennas. This has a massive effect on bandwidth efficiency and has another advantage in that there is less effect if an antenna becomes inoperative,” Yost explained.
With this proliferation of antennas, Massive MIMO offers full spatial multiplexing where multiple data streams are transmitted at the same time and on the same radio channel and is a crucial part of 5G network development. It is proving to be a successful path to follow as spectral efficiency improvement multiples of at least 10 times have already been achieved.
Yost told me, “Spectral efficiency is measured in data per second per Hertz. A figure of 5-10 bits/s/Hz is normal for 4G networks but our 5G experiment using full spatial multiplexing has already achieved a record figure of 145 bits/s/Hz”
Commenting on the work carried out on Massive MIMO at BT’s Adastral Park facility, the company’s Managing Director of Research and Innovation, Professor Tim Whitley stated, “Massive MIMO has the potential to significantly boost available data rates in future 5G mobile networks.” As such, Massive MIMO is good for enhanced broadband and goes some way towards fulfilling the needs of the IoT but it doesn’t address the problem of available bandwidth.
With the sub-6GHz part of the spectrum being crowded with small chunks of it selling for millions of dollars, there is a pressing need to explore areas of the radio spectrum that are currently unused commercially.
One frequency range that has large amounts of availability is the so-called millimetre wave band between about 30GHz and 300GHz. Some of this is used by military organisations and is out of bounds but there are still large tracts of unclaimed bandwidth waiting to be used. However, the challenges of propagating that bandwidth are significant.
“The problem with mmWave radio is that it doesn’t go through walls and it’s heavily attenuated by water,” Yost explained. For this reason, National Instruments is working with AT&T on channel sounding to understand the effects of different environments, and with NYU on beam steering to look at overcoming the barriers of different transmission media.
To further this research, NI recently provided NYU with hardware and software from its SDR range. Such mmWave SDRs and the associated test and measurement equipment are necessary to transition mmWave technology from concept and simulation in the lab to a real-world environment.
According to James Kimery, director of RF research and SDR marketing at NI, the purpose of the donation was to help advance this research to solve the challenging problems the industry faces migrating to mmWave for 5G.
Commenting on the development work being done at NYU, Professor of electrical and computer engineering, Sundeep Rangan said, “mmWave wireless prototyping demands platforms with enormous baseband processing power along with advanced antenna array systems, which are extremely difficult to develop in university labs. With NI’s SDR equipment, we’re now able to perform rapid prototyping and experimentation to push the envelope in mmWave channel sounding, emulation and communication system design and drive the development and commercialisation of mmWave technology.”
With such innovative research taking place on both Massive MIMO and mmWave, there is still the challenge of timeframe with the IoT looming and potentially demanding more bandwidth and efficiency than is available. I asked Yost for her view on this.
“To overcome some of the timeframe problems, it will be possible to look at different elements of the network in which the technology can be deployed, such as using mmWave for fixed wireless access. It is entirely feasible to have mmWave available for fixed wireless access to address the “last mile” challenge but it is unlikely that it could be used for consumer handsets and IoT endpoints, which will still be using 4G technology,” she explained.
She went on to say that we’re just scratching the tip of the iceberg at the moment with the IoT and that it isn’t just the physical layer, but also the MAC layer that needs addressing. The MAC (Media Access Control) layer is where the communication protocols are defined that enable the signals that are sent on the physical layer to be handled and managed correctly.
“The Massive MIMO and mmWave research that is going on is all at the physical layer and there is still a lot that needs to be done in this respect. Bristol is looking at spectral efficiency and Lund is examining calibration and the ability to use it commercially,” she said.
I asked Yost if it was likely that 5G availability would be a constraint on the deployment of the Industrial IoT and she told me, “I think they’ll develop in parallel. A lot of the impetus comes from the network providers. The IoT has a different pricing structure to mobile communications. Data usage is small per device so the pricing model has to change. Once they realise the potential, there will be more focus on providing the infrastructure.”
Bucking the trend
The historic cycle of 10 years between each generation could be forced to change with the demands of the IoT. Since 2025 is five years too late for meeting the needs of both the commercial and industrial internets of things, the generation gap is likely to be bridged by a more blended evolutionary process.