Researchers at the University of Technology Sydney are at the cutting edge of research in mobile communications technologies that will boost the capacity of future networks by making much more efficient use of spectrum and by enabling very high frequencies, in excess of 100GHz, to be used for mobile communications.
The next generation of cellular technology, 5G, is being developed and deployed at breakneck speed to meet demand for bandwidth data volumes and device numbers that are all increasing at exponential rates.
However, the technologies being incorporated in the 5G standards, expected to be finalised in 2018 will not meet surging demand indefinitely. Mobile communications at frequencies much higher than those in use today, or envisaged in the 5G standards, and much more efficient ways of using existing spectrum are already being explored.
Professor Eryk Dutkiewicz, head of the School of Computing and Communications in the Faculty of Engineering and IT at UTS, leads a team working on 5G-related technologies.
An industry focussed research facility
Dutkiewicz started work on a 5G related research project funded by Intel at Macquarie University but moved the team to UTS, which he described as being more industry friendly. He said his team at UTS was very much industry focussed, with much of the unit’s funding coming from industry backers outside Australia.
“We are seen as a very unique facility and a capability that can be tapped by industry. We have range and depth. Xiaojing and I [Xiaojing Huang, another member of the team] used to work for Motorola. We know how industry works and how to work with industry.
“We are not just a group of academics who want to write papers. Our supporters are not interested in publications per se, but in things that can go into their product roadmaps and into their intellectual property.”
Dutkiewicz’s team sits within the Global Big Data Technologies Centre (GBDTC), an international centre of excellence developing technologies for big data science and engineering, working closely with industry
The Intel funded research is looking at ways to comply with requirements already being demanded to address spectrum shortages in the US and Europe.
“In the US and Europe spectrum is a big problem right now, less so in Australia,” Dutkiewicz told Computerworld. “Our work with Intel is funded by Intel Europe and US for solving European and US problems, but they picked us because we have the expertise to do it.”
Dutkiewicz explained that both the FCC in the US and CEPT the Europe are moving to allow other users to have access to spectrum presently licenced for the exclusive use of the incumbent.
Spectrum too precious for exclusivity
The first spectrum to which this regime will be applied in the US is in the 3.5GHz band, used by the US Navy for radar. “The FCC is proposing three tiers of users: the incumbent; somebody who is a licensee and will pay for usage; and somebody who will have opportunistic access without having to pay for it, but they must still be licenced,” Dutkiewicz said.
CEPT is proposing just two tiers: the licensed incumbent and a secondary licensee, and is planning to apply this regime to frequencies in the 2.3GHz band used for wireless news-gathering in some countries, and in France for air traffic control radars.
“This standard for sharing it is called the Spectrum Access System in the US and Licensed Spectrum Access in Europe,” Dutkiewicz said.
Strict controls will be needed to ensure that the secondary usage does not impact services operated by the licensed incumbent. Current plans are for secondary usage rights to be granted for periods of several months, Dutkiewicz said, but his team at UTS is developing technologies that could allow secondary usage to take advantage of unused spectrum for period that could be as short as a few seconds.
“If you look closely at the spectrum you will find lots of ‘holes’ on a very small time scale,” Dutkiewicz said. “We are building a demonstrator that will be able to sense the spectrum, create radio frequency maps in very fine granularity of time, recognise the opportunities and take advantage of them. We are trying to determine the optimal time scale. Maybe it is seconds, or milliseconds.
“We want to see how fine a map we can build, what are the computational requirement behind it, what are the implications of measurement errors. We have done some theoretical work, now we want to test it.”
He said the team hoped to have the demonstrator in operation late this year or early next year.
Spectrum usage info for sale
He envisages making maximum use of spectrum becoming such a priority in the future that users’ devices will be co-opted into providing information to identify ‘holes’ of unused spectrum. Also, he says the need to maximise spectrum usage could see the emergence of specialist monitoring agencies operating networks of sensors and on-selling spectrum usage information to those wanting access to the spectrum.
“Spectrum knowledge will be a very valuable commodity in the future if the paradigm changes from owning the licence to sharing the licence,” Dutkiewicz said. “Someone independent of the mobile operators could build a sensor network to provide a service gathering information about spectrum, processing that information and selling it to network operators. That would be very useful information.”
Another possibility is co-opting user devices to provide information on spectrum availability. “Alternatively, operators could use their customers to provide them with spectrum information,” Dutkiewicz said. “Maybe in a future iPhone there will be a smart spectrum sensor and an app. If you switch this app on there will be some advantage for you because you are helping the operator get spectrum information.”
Dutkiewicz said this functionality was not under consideration for the upcoming 5G standard, “but we are working on it.”
Accessing untapped spectrum
Spectrum sharing techniques will enable more traffic to be carried on the spectrum already in use today, but the researchers at UTS are also working on antenna technologies for frequencies much higher than those in use today. These signals would have very short range, but would enable data rates of many gigabits per second that could open up a whole range of new applications.
Rick Ziolkowski, who has recently joined the team from the University of Arizona, is working on antenna designs for these very high frequencies. “I know a group in Tokyo that is looking at the idea of you downloading a movie in the time it takes you to walk through security,” he said.
He told Computerworld that UTS researchers were trying to develop antennas for terahertz frequencies “We will be trying to build the antennas and test them: not just do theoretical modelling it will be build and test.”
Antenna technology at these frequencies is totally different to that used for today’s cellular networks. Antenna size is determined by wavelength: the higher the frequency the shorter the wavelength and the smaller the antenna needed.
At 3.5GHz, the highest frequency used in Australian mobile networks today, antennas are a few centrimetres in size. At 100GHz they would be just a few millimetres, and at terahertz frequencies less than a millimetre.
“We are trying to develop smaller antennas, to understand how to avoid mutual coupling when you get a lot of antennas close together, and how can they be made configurable so they can be used with different frequencies, different polarisations,” Ziolkowski said.