What is 5G? – that is what the Brooklyn 5G Summit is trying to answer by bringing together many of the leading companies and researchers in the wireless communications industry to share their latest research results and thoughts on the subject. From the semiconductor level (like Intel, Qualcomm and UC San Diego) to channel modeling (like NYU WIRELESS, University of Bristol, Nokia and Ericsson) to network providers (like China Mobile and AT&T), the whole mobile communications ecosystem was covered at this first time event. Professor Ted Rappaport, who formed NYU WIRELESS which has quickly become a leader in mmWave communications research, hosted this first of its kind event along with Nokia Solutions and Networks (and the NYU WIRELESS affiliate companies).
5G standards are expected to be addressed in release 15 or 16 of the 3GPP standards around 2018-19 targeting systems in the 2020 time frame (a new generation of mobile communications systems is released about every 10 years or so). Researchers are now developing various technologies that could be used in 5G systems but nothing is determined at this point in time. Some likely technologies include mmWaves, massive MIMO (antenna systems using hundreds of elements in an array), software defined radios, cloud-based networks, self-organized cells, alternative modulation schemes to OFDM, device to device communications and others.
The first key thing I learned is that mmWaves (work was presented at various frequencies from 28 to 90 GHz) do not exhibit the expected problem of signal loss versus frequency (from oxygen absorption in air) due to the positive effects of diffuse scattering in many environments. The non-line of sight (NLOS) signals can be gathered from multiple beams as they are scattered from the environment gaining signal strength. NYU WIRELESS testing has shown that at least 3 significant beam angles are available in link testing done in NYC with up to 5 in some cases. The higher frequencies are very desirable for the obvious reasons of reducing antenna sizes for large arrays, wider bandwidth communications, more available spectrum, less interference (more directive signals), etc.
One example of mmWave technology already in use is for 60 GHz WLAN applications. The WiGig and WiFi Alliances have merged leading to a standard for 60 GHz WLAN and WPAN. Intel is integrating WiGig into new PCs using 60 GHz multi-beam modules they have developed with 12 elements. The WiGig 2 standard is already being worked on with over 50 Gbps data rates expected. The market projected by some analysts is over $10 billion by 2019 for this type of technology. Intel has demonstrated that this technology can be used for cellular systems and short range backhaul. Scaled 128 element systems could be used to 400 m for LOS systems and 100 m for NLOS at 1 Gbps speeds.
Also on the semiconductor side, Professor James Buckwalter of UC San Diego discussed how CMOS has progressed to a point where low cost PA arrays can provide greater than 1 W at E-band with an EVM of less than 1.2% (1024 QAM). CMOS FETS can be stacked to increase the breakdown voltage to achieve high performance PAs and outphasing techniques can greatly improve the efficiency (plus DPD is used to lower distortion). He gave 3 specific examples including a 45 GHz device (45 nm CMOS SOI) with a max output power of 18.6 dBm, PAE of 34% and by using 8 way power combing to 4 antennas, greater than 30 dBm was achieved providing 1.3% EVM (1024 QAM with DPD) and 1 Gbps data rate. Second was a 21 dBm max output power, 16% PAE device with greater than 1.2 Gbps data rate (QPSK). Last was a 23 dBm output power device (20 dBm per channel) with greater than 60 dB dynamic range, EVM of 2.4% (256 QAM) with greater than 1 Gbps data rate probably representing the fastest microwave outphasing device to date.
Dr. Andrea Goldsmith of Stanford University discussed how the FCC stated that even if we used the entire available spectrum that is currently allocated in the US today, we would be 300 MHz short for the projected needed capacity (although WiFi offload is not accounted for in this estimate). She emphasized how we do not know the Shannon capacity of wireless channels so we are not sure of our limits. She covered how traditional cellular systems are interference limited but small cells will exponentially increase capacity and would reduce the effects of interference (they eliminate pilot contamination that limits capacity). Using mmWaves would also reduce the interference issues and other alternatives such as non-coherent networks could work at mmWaves. Software defined networks are also a key for future systems and the Internet of Things (IoT).
Ted Rappaport discussed his work with mmWave channel measurements done at 28 and 73 GHz. His results show that the impression that mmWaves will be highly absorbed in air so are not effective for mobile wireless systems is a myth. The absorption is typically less than a dB/km so not that dramatic (except perhaps at 60 GHz where there is a large peak in absorption). Rain is also not a huge impact and typically only reduces the signal by a few dB. His results show that 3GPP and WINNER models currently used do not work for mmWaves so new models need to be developed. As stated, he has found that at least 3 angles of energy arrival are available to extend the range of mmWave systems in his measurements made around NYC. There is little diffraction effect (like in lower frequency systems) and more scattering effects that add gain to the signals.
Samsung announced in May of 2013 that it had successfully developed the world’s first adaptive array transceiver technology operating in the millimeter-wave Ka-bands for cellular communications. Samsung’s new adaptive array transceiver technology has proved itself as a successful solution according to the release so Dr. Wonil Roh discussed how the system uses 28 GHz to transmit data at a speed of up to 1.056 Gbps to a distance of up to 2 km. The adaptive array transceiver technology, using 64 antenna elements, could be a viable solution for 5G systems and is viable for handsets according to Samsung testing.
Dr. Thomas Haustein of Fraunhofer HHI talked about MiWEBA which is a joint European and Japanese research activity targeting enhancements of mmWave technology for mobile backhaul and access application scenarios. They believe an overlay of mmWave cells on top of the current cellular network would address the capacity issues projected. This approach involves two primary methods to increase network capacity: network densification and spectrum extension. In contrast to current approaches, that aim for increasing the capacity by adding smaller cells operating in the same frequency band and hence run into interference limitations, no interference is induced by the mmWave overlay. They advocate splitting the user- and control-plane in order to have mmWave small cells integrated in an effective and efficient way keeping the control plan constantly attached to the classical sub- 6 GHz IMT bands while the user plane can be switched between IMT bands and mmWaves.
China Mobile Research Institute covered their strategy for 5G that emphasizes “Green” and “Soft” as their cornerstones. Dr. Chih-Lin I discussed the need to reduce energy consumption and use software defined radio networks built using cloud radio access networks (C-RAN). With 800k users (several time more than the than the entire US network providers combined), she said we need to re-think Shannon/Ring & Young and come up with new strategies. She envisions systems that optimize energy efficiency and capacity dynamically (no more cells) using C-RAN systems. She also emphasizes invisible base stations such as using smart tiles on the sides of buildings that blend into the architecture. From their talk at EDI CON 2014 in Beijing, we learned that the largest power consumption comes from the air conditioning needed to cool the base stations so they plan to virtualize these into the cloud and congregate them together to reduce energy needs.
Tommi Jämsä of Anite talked about the European METIS (Mobile and wireless communications Enablers for Twenty-twenty Information Society - good thing they came up with an acronym) project that investigates the next generation radio technologies. Since current models are not adequate, METIS is coming up with channel model requirements, scenarios, a framework and measurement results about the first METIS channel model soon to be published.
The event included exhibits by supporting companies outside of the auditorium with table top displays. National Instruments is an affiliate and there equipment is used by the NYU WIRELESS students in their channel sounding measurement system. NI had on display their rapid proto-typing software defined radio system with 1 GHz bandwidth. Their systems are modular, flexible and use LABVIEW so can be quickly customized and setup for measurements on almost any kind of communications systems. Rohde & Schwarz had on display their 67 GHz VNA for making mmWave measurements and demonstrated how components can be characterized using their system. They had a waveguide filter being tested showing the various measurements that can be made to characterize its specific performance parameters which can vary with high frequency components. Agilent was also on site showing off their VNA capabilities and other testing capabilities. They had a SiBeam 60 GHz device doing live video streaming and showed the various measurements that can be made including how the reflected mmWave signals behave bouncing off of an object and how directional they are so that blocking can become an issue. View our video gallery for demos from these 3 companies.
Professor Rappaport and NSN ran a great inaugural event that attracted the best minds in the industry and academic world doing advanced research on potential 5G technologies. It is surely going to be a high level annual event for many years to come as these technologies are tested and verified eventually leading to a standard.