With the first commercial launches of 5G expected within five years, research into 5G systems is revving up with partnerships and exploratory steps around the globe.
RCR Wireless News spoke with Tod Sizer II, who leads Alcatel-Lucent Bell Labs’ wireless research program and is in charge of research teams in seven locations around the world that focus on wireless systems, software and technology, as well as convergence with fixed-line systems. Sizer gave us his top 10 areas for 5G research:
10. Massive MIMO. Massive MIMO goes far beyond the simple 2×2 or 4×2 MIMO that is in use in some networks today, to arrays with dozens or 100 antennas. Sizer said Bell Labs has developed a 64-antenna array, but there’s still considerable work that needs to be done to make sure arrays are cost-effective and deployable. They would be more likely, due to size and weight, to end up on the side of a building than on a cell tower, he said.
NI has documented its support of researchers’ capabilities for building massive MIMO test beds in a new white paper published this week, with one example at Lund University in Sweden.
9. Centimeter and millimeter wave. This should come as no surprise to anyone who’s been paying attention to 5G chatter. There is a substantial amount of spectrum at very high frequencies, which makes it very attractive, but the engineering challenges are intense. Sizer said that this spectrum offers “huge opportunities and huge challenges,” with the latter including the spectrum’s vulnerability to shadowing.
“If there is not a line-of-sight link between the access point and the user device, then the connection basically goes to zero – unless you’re lucky enough to have a reflection off a very flat wall nearby,” Sizer explained. He said that researchers and engineers will have to figure out how to leverage that spectrum while providing the consistent high quality of experience, adding that performance over distance will be important as well. Sizer said he expects systems that can harness such spectrum over a meter or two of distance from the base station will start to appear soon, such as WiGig, but it will be important to design systems that can utilize the high-frequency spectrum at a distance of, say, 100 meters.
Work is already ongoing to explore channel models for 5G spectrum. You can watch a moderated panel on the technical challenges of 5G here, and check out this interview with Wilhelm Keusgen of the Fraunhofer Heinrich Hertz Institute about his research in this area.
8. Multi-Radio Access technologies. Sizer said that from Alcatel-Lucent’s perspective, 5G is likely to involve Wi-Fi, traditional cellular frequencies and high-frequency spectrum. The ability of devices to talk to different types of access points at the same time is already present in networks to some extent, he pointed out, with LTE and Wi-Fi.
“This is the way we think we’ll be able to support centimeter and millimeter wave,” Sizer said; a highly reliable connection will be maintained on a traditional cellular frequency, with surges of extra speed on millimeter wave.
7. Battery life. This one may not be as obvious, but Sizer said that figuring out new ways to extend battery life will be extremely important – particularly for the “Internet of Things” ecosystem, where sensors may need batteries that can last a decade or more and be produced economically.
Mike Wright, director of Telstra’s networks, has said that more efficient networking in 5G could be a component of improving battery life, which has been hit by the popularity of devices with large screens.
6. Research into a new wave form, or air interface. Although some have posited that 5G may not involve a new air interface, Sizer pointed out that a fundamental component of a new generation of technology is the point at which it is no longer backwards-compatible with the old. Sizer noted that the messaging between the network and the device in LTE eats up a huge amount of battery life and network resources even when the related information is minimal, and that for applications such as car-to-car communications, tactile Internet and other potential use cases that require sub-10-millisecond latency, a baseline change in the technology will be needed.
“There are lots of improvements we’re obviously going to be able to make to LTE, but there are some that we believe will require a more fundamental change to happen,” Sizer said.
5. Ensuring end-to-end performance. The common denominator of future networks, Sizer said, will be that they will all be “untethered,” or wireless for at least the last few feet. Meanwhile, users “will have expectations of unlimited capacity and they will have expectations of infinite response” – essentially, the feeling that the network was built for and serves them alone. Although there’s no such thing as infinite capacity, Sizer said that more intelligent and efficient networks should be able to assess what its users are doing and allocate resources intelligently according to application. This will require better network coordination and some innovation in the Radio Access Network, he said, as well as the core network, content delivery networks and backhaul.
4. Contextual awareness. This goes along with end-to-end performance and is the ability for a network to understand what device a user is on, what application is being used, the physical location and speed, and adapting the network performance to best serve those parameters. Work in that area is already ongoing.
3. Intelligent data mining on the fly. Big data analytics is already an area of interest for telecom, and Sizer expects data-based intelligence to become more prevalent as part of contextual awareness. However, he acknowledged that as more and more content providers move to encrypted content, network providers have less visibility into those data streams. This is a very active area of research, he said, and much can be inferred about the content of a data stream based on its behavior. Sizer said he expects to see content providers and network operators come around to sharing more information with one another, because both sides ultimately want an excellent end-user experience.
2. The cloud – particularly, the distributed cloud – software-defined networking and network function virtualization. As the industry explores more flexible, automated network solutions, this part of the evolution toward 5G capabilities is already underway and Sizer expects it to be fundamental to 5G. However, he pointed out that research questions remain on the best network architectures for different applications. Ultra-low-latency applications such as autonomous driving may require highly distributed networks simply due to the laws of physics, while applications that can tolerate higher latency could be served from fewer central locations.
1. New applications and requirements that can leverage and take advantage of an “always connected, always untethered, infinite capacity and response world” that Sizer describes. If there is no business case for 5G, it won’t come to fruition.