5G: The Next Generation Mobile Network
A Brief History of the Evolution of Cellular Systems
The first generation (1G) system was an analogue system in the 1980s. Its purpose was to provide basic mobile voice service using FM technology, as in FM radio. Mobile phones resembled bricks and voice capacity was very limited.
The second generation (2G) of digital cellular systems, from the end of the 1990s to the early 2000s provided much higher voice capacity. Since voice capacity limitations had been solved, there was a low subscription fee. However, 2G was expensive as it charged by airtime.
The continued evolution from 2G to 3G was not motivated by voice capacity, but by the need to incorporate wireless mobile internet access. Many consumers were skeptical of why they would want internet on their phones; clearly this is no longer the case.
With the popularity of smartphones, the need to make data application faster pushed the development of 4G, which boasts a much higher bit rate than 3G. In many countries, 4G is expensive and still isn’t widely used, so why the rush for 5G?
Three Reasons for 5G
Lau identifies three major perspectives as to what’ll set 5G ahead of 4G.
• Speed - The first perspective is concerned with making communications 1,000 times faster. Over the past 10 to 20 years, capacity demand has been exponentially increasing. It is projected that by 2020, capacity demand will be 1,000 times higher than it is today.
• Scale - The second perspective considers the need to allow 1,000 times more devices. Today, most communications involves people to people or people to machine communications. With the advent of the Internet of Things, machine to machine communication is growing. This means that in the future, trillions of devices will need to be interconnected - and with short latency, as some sensory applications will need to send short-bursts of data faster than any applications to-date.
• Control - The third perspective considers the control aspect of a network. Currently we have a tactile network for communications and the shortest latency possible is around 50 milliseconds. This allows us to have real-time conversations over applications such as Skype, which use voice over IP. In the future, applications will go beyond communication. There are presently plenty of scenarios for machine to machine communications in industrial processes. Sensors collect state information of potentially unstable dynamic plants and communicate to remote controller or actuators for stabilisation of the entire process. However, for smooth control, latency less than 1 millisecond is required. Therefore today’s latency standard is too slow; it’s quick enough for communication but not for accurate control.
The Three Stages of Mobile Network Development
Stage One is theoretical and involves the research community. It is the stage at which industries present market demands and universities and industrial researchers consider how their research may facilitate these demands. At this point, researchers figure whether it is possible for technology to meet these demands. It might also involve proof of concept testing where researchers try to build prototypes and testbeds to verify if an idea or algorithm will work in the field.
Stage Two involves the establishment of industry standards for the new generation system. These standards are set by the International Telecommunication Union to ensure that while companies have enough flexibility to innovate and compete, technologies are compatible because they conform to a common standard.
Stage Three involves implementation and deployment of the system. After these standards have been agreed on, different vendors produce the infrastructure and the handset and different operators lay out the networks.
The Next Generation Mobile Alliance estimates that 5G will become available in 2020. Lau estimates that at present 5G development lies somewhere between the first and second stages.
The Challenges Facing 5G Development
There are a number of challenges to overcome in mobile network development. The first key issue is that at present, to connect to 4G, the device has to connect to the base station before sending data packets. The protocol involved is inefficient and heavy if the device just needs to send a small burst of data. With the predicted 1,000 times more data to process and 1,000 times more devices to connect, it will be important to streamline in order to fit small bursts of urgent data, with minimum overheads.
The second key issue concerns the limitations of radio bandwidth. Because of this, license fees are expensive and the valuable spectrum has to be reused in a cellular network, causing interference. To accommodate the projected 1,000 times’ capacity and latency, interference issues must be addressed in a fundamental way. There are two ways that new technologies can help:
1. By using the existing spectrum more efficiently. If capacity is to be increased by 1,000 times, bandwidth cannot; so the question is how to squeeze in 1,000 times more customers. One proposed solution is to get technology to use the existing spectrum more efficiently. Lau is working a technology with great potential for this, called the Massive MIMO (multiple input multiple output).
2. Another solution is to use a different spectrum such as millimetre wave. Going to higher spectrum opens up more bandwidth but poses challenges in the propagation physics. For instance, at higher frequencies, radio waves tend to suffer from larger attenuation and propagate based on line-of-sight, meaning that the device has to see the transmitter in order to receive the signal. Massive MIMO is also useful to overcome the limitation of the poor propagation and therefore, fully unleashing the huge bandwidth available in these high frequencies.
What is Massive MIMO?
Multiple-antenna technology is one of the greatest inventions of wireless communications in the last 20 years and is already widely used in various communication standards such as 3G, 4G, and WiFi. By exploiting the scattering in the radio channel, it can achieve higher spectral efficiency and higher reliability. With Massive MIMO, we are talking about hundreds of antennas at the transmitter and/or the receiver. This massive number of antennas allow order-wise higher bit rates over limited spectra as well as higher spatial freedom to mitigate interference in the network. Therefore, it is a promising way to realise the challenging goal of 5G systems. However, the massive number of antennas also pose a lot of implementation challenges such as the complexity of the transmitter and receiver, the signaling overhead between the transmitter and receiver to estimate the channel as well as the accurate tracking of spatial beams. Lau’s research is focused on low complexity architecture and algorithms to fully unleash the potential of massive MIMO to future 5G systems.
Vincent Lau has a BEng (1992) from the department of Electrical and Electronic Engineering, University of Hong Kong. Following graduation, Lau joined Hong Kong Telecom as a systems engineer for three years. Lau completed his PhD (1997) at the University of Cambridge with the support of the Croucher Foundation, the Sir Edward Youde Memorial Fellowship and the Rotoract Scholarship. In 1997, he joined Bells Labs - Lucent Technologies, New Jersey as a Member of Technical Staff. In 2004, Lau joined the department of Electronic and Computer Engineering, Hong Kong University of Science and Technology as an Associate Professor. Since 2014, Lau has been Chair Professor of the department. Lau has also acted as technology advisor and consultant to a number of companies such as ZTE, Huawei, ASTRI, B3G, WiMAX and Cognitive Radio. He is the founder and co-director of Huawei-HKUST Innovation Lab.
To view Vincent's personal Croucher profile, click here.