Better 5G with AI?
November 29th, 2018
By Lynnette Reese, Editor-in-Chief, Embedded Systems Engineering
The long-promised 5G cellular network is slowly rolling out, one stage of standards at a time, with much marketing fanfare. The final and complete global standards for 5G are to be released in 2020. A patchwork of different communications technologies will combine ultra-densification, massive multiple-input multiple-output (MIMO), and use of the mmWave spectrum to achieve the demands of the wireless data explosion that is already upon us, for which 5G will be the answer.i
The path of innovation has often started with the launch of new technology, followed by new applications, then by consumer or industry demand, and results in commercial or personal productivity. Today, the tail is wagging the dog in that 4G is not ready to support the rapidly growing and extremes in demand, i.e., the technology to sustain enormous flows of data via 4G cellular technology is not here yet.
However, rescue comes as 5G, which presents a 100x to 1000x improvement over 4G, and is expected to reduce latency further, increase throughput by 1000 times (as an aggregation of a patchwork of technologies), and improve coverage to 100 Mbps or more for at least 95% of users. Such an enormous improvement relies on several key technologies to achieve offloading and extreme “densification” since the physical frequency spectrum is not able to expand. (Technology can improve the efficiency of the spectrum, however.) Engineers see a possible increase in bandwidth by expanding into the mm wave spectrum. We can make better use of the unlicensed spectrum and apply MIMO to sustain even more bits/s/Hz per node.
5G will service a diversity of needs such as delivering live streaming HD video to large numbers of mobile devices, real-time two-way mobile gaming, touch-screen activated cloud-based applications, a trillion or so IoT devices that send data 24/7, and autonomous vehicles with ultra-reliable, low latency service for Vehicle-to-Everything (V2X) communications.
Advanced 5G networks, using AI, could quickly determine variations in network traffic, adjust resources based on demand, coordinate base stations and other networked equipment, and use elementary intelligence to provide options for diverse situations.
According to Intelligent 5g: When Cellular Networks Meet Artificial Intelligence, “5G will need to be a paradigm shift that includes very high carrier frequencies with massive bandwidths, extreme base station and device densities and unprecedented numbers of antennas…the core network will also have to reach unprecedented levels of flexibility and intelligence, spectrum regulation will need to be rethought and improved, and energy and cost efficiencies will become even more critical considerations.”ii
It seems as if the decade-long sustained effort that we will have put into the logistics, design, coordination, deployment, and launch of 5G is the kind of effort, diligence, and commitment needed to fix something like a broken national political system. For engineers, math is either right or wrong; large decisions are based on physical truth, correct mathematical calculations, and some experimentation. Alas, “necessity is the mother of invention,” and human nature is to wait until a system is untenable before working on a solution.
[i] What Will 5G Be? Jeffrey G. Andrews, et al., 12 May 2014, https://arxiv.org/pdf/1405.2957.pdf
[ii] R. Li et al., “Intelligent 5G: When Cellular Networks Meet Artificial Intelligence,” in IEEE Wireless Communications, vol. 24, no. 5, pp. 175-183, October 2017.
Lynnette Reese is Editor-in-Chief, Embedded Intel Solutions and Embedded Systems Engineering, and has been working in various roles as an electrical engineer for over two decades.