Industrial Wireless Part 2: Forget the Hype – What Can 4G LTE and 5G Do Today?
| By: Kristin Lewotsky, Contributing Editor
Editor’s Note: In part one of this three-part article, we covered industrial wireless communications over unlicensed spectrum. In part two, this article, we will focus on the current state of 4G LTE and 5G cellular wireless and the technologies that make them effective for industrial applications. In part three, we will bring it all together to show how to choose the ideal industrial wireless solution and how to efficiently and effectively deploy it.
As the march toward flexible manufacturing, smart factories, and digitalization continue, connectivity is an ever-growing need in industrial automation. For many years, wired connections ruled the day. As industrial users have begun connecting more and more devices and personnel to their networks, however, wireless networking increasingly offers a faster, easier, and potentially more reliable solution than running cable. “The LAN cable won’t go away but there are challenges just in terms of cost when it comes to trying new cables to every single device that now needs to be connected,” says Stephane Daeuble, head of enterprise solution marketing at Nokia. And in the case of mobile applications, of course, it’s the only game in town.
The latest releases of cellular technology, 4G LTE Advanced Pro and 5G, have been engineered to deliver high data rates, predictable low latency operation, and broad coverage. Because they are purpose-built for mobile applications, they offer a degree of reliability and security not available with the alternatives. As the latest generation of wireless, 5G dominates the headlines with predictions of world-changing performance. While it is certainly likely that many use cases will come to fruition, the rollout of any wireless generation is an iterative process. Although 5G’s high-speed capabilities are in full view, many of the game-changing functionalities being touted will not reach the market for several years. In the interim, organizations should be looking at hybrid solutions of the latest generation that goes beyond LTE Advanced Pro (known colloquially as 4.9G) and the currently available versions of 5G.
Key technologies for 4G and 5G
A new generation of cellular wireless is rolled out about every 10 years in a multiphase process orchestrated by the 3rd Generation Partnership Project (3GPP). 3GPP releases are developed, finalized, then codified as standards by telecommunications standards organizations. In the 42 years since the release of what is now known as 1G cellular, data rates have advanced from a few kilobits per second to 5G’s promised peak download speed of 10 Gbps (see table 1).
In a wireless link, a narrowband channel in a frequency band transmits a modulated carrier wave that sends the signal between a base station and user equipment (UE). Although 3G and even 2G cellular networks are fast enough to support limited industrial Internet of things (IIoT) sensor networks, cellular technology didn’t really come into its own for the industrial environment until the emergence of 4G LTE. To boost data rate and coverage, 4G LTE includes several modulation and transmission technologies (see figure 1). Many of these technologies have been modified and carried over to the 5G standard.
|Figure 1: In this representation of cellular air interface and modulation technologies, passengers represent data and the colors of the vehicles represent different user devices. (Courtesy of Sierra Wireless)|
Quadrature amplitude modulation (QAM)
QAM is a complex modulation scheme capable of encoding multiple bits per symbol using multilevel amplitude modulation on a pair of carriers that are 90° out of phase (in quadrature) with one another. 16-QAM, for example, uses four amplitude levels instead of the two levels used for conventional binary modulation; as a result, 16-QAM can achieve 16 distinct phase/amplitude states, enabling it to encode four bits per symbol. 4G LTE debuted with 64-QAM (six bits per symbol) on the downlink. The most recent release, LTE-A Pro, uses 256-QAM (eight bits per symbol) on both downlink and uplink.
Orthogonal frequency-division multiplexing (OFDM)
In OFDM, the channel is divided into a set of independently modulated subcarriers with frequencies that are orthogonal to one another (i.e., the peak of one subcarrier wave aligns with the nulls of the others’ frequency spectrum to prevent interference). During transmission, the bit string is divided into multiple substrings that are then simultaneously encoded onto a set of subcarriers. Transmission takes place at slightly slower data rates but the fact that the bit string is multiplexed means that the overall transmission is faster.
The previous two techniques describe ways to encode more data onto a carrier wave. 4G LTE also applies techniques to send more data on a single frequency band.
Carrier aggregation is a technique for combining multiple carriers (component carriers) at different frequencies into a single “supercarrier” that will be sent to the same device (see figure 2). Aggregating five 20-MHz component carriers, for example, yields a total bandwidth of 100 MHz to a single device.
|Figure 2: In carrier aggregation, component carriers of identical or different bandwidths are combined into a single “supercarrier” in order to increase data sent over the same frequency band. (Courtesy of 3GPP)|
Introduced as part of LTE Advanced, carrier aggregation initially allowed a maximum of five component carriers of identical width. LTE Advanced Pro supports 32 component carriers of mixed channel width. Carrier aggregation works best when applied to different channels within the same overall frequency band but can be used with component carriers of different frequencies.
Multiple input/multiple output (MIMO)
MIMO is a spatial multiplexing technique that takes advantage of differences in the propagation paths of carriers to enable more data to be sent simultaneously. Instead of using a single antenna to send a single transmission across a single physical path to a single receiver in the UE, MIMO uses multiple antennas to simultaneously send multiple carriers to multiple antennas within the UE (see figure 3). Because the propagation paths are slightly different, the signals do not interfere. Single-user MIMO (SU-MIMO) is primarily used in 4G LTE Advanced Pro as a way to increase the overall data rate.
|Figure 3: Multiple in, multiple out (MIMO) transmission uses multiple antennas at the base station and device to increase overall capacity. This diagram shows a 4x4 MIMO system, as introduced in 4G LTE. (Courtesy of Sierra Wireless)|
The RF band is broken into licensed and unlicensed spectrum. Licensed spectrum is controlled by national regulatory bodies and is assigned to specific applications or clients such as critical infrastructure, public services, vertical industries, or paid licensees like communications service providers (CSPs) and private enterprises.
Cellular networks have traditionally operated in the licensed frequency bands (see table 2), while local-area networking technologies like Bluetooth and Wi-Fi operate in the unlicensed bands (900 MHz, 2.4 GHz, 5.8 GHz, 6.0 GHz).
Release 13 of 4G LTE introduced Licensed Assisted Access (LAA), in which networks can augment downlink capacity with unlicensed spectrum using carrier aggregation. The technique involves using the core infrastructure and licensed carriers as “anchor bands” that handle control-plane communications like call initiation and termination, authentication, security, etc. The unlicensed spectrum can only be used for overflow user-plane communications (data and voice). To prevent cellular transmission from interfering with Wi-Fi networks that traditionally operate in this spectrum, the standard mandates use of a technique called listen before talk (LBT) to seek out clear channels. When channels are in use, frameworks have been put in place to try to ensure fair use between cellular and Wi-Fi, although a certain amount of controversy does exist.
Release 13 added another option, LTE Unlicensed also specified for the 5 GHz band. LTE-U still requires an LTE anchor channel but uses a purely unlicensed spectrum for user-plane communications.
MulteFire, defined by the MLFA, created a version of LTE that can also operate fully in unlicensed spectrum, like Wi-Fi.
Designed to address the explosive growth of mobile applications and the rise of the Internet of Things, 5G was crafted to provide three key service levels:
- Enhanced mobile broadband (eMBB)—downlink speeds as high as 10 Gbps Applications include augmented reality/virtual reality; high-speed video, for example, to support high-resolution inspection.
- Massive machine-type communications (mMTC) —1 million devices per square kilometer mMTC supports IIoT for use cases like digital twins or predictive maintenance.
- Ultrareliable low-latency communications (URLLC) —1 ms latencies with 99.9999% reliability
- Applications include autonomous vehicles and remote control of highly coordinated high-speed equipment.
Before diving into a discussion of the 5G New Radio (NR) and the other modifications that make this performance possible, it’s important to make a couple of observations. First, the numbers above are theoretical. Actual performance, particularly initially, will be considerably more modest (see table 3). Particularly in the case of 5G, real-world speed will be affected by everything from carrier bandwidth to transmission frequency.
Performance is also affected by the second point, which is that, like all wireless standards, 5G is rolling out in phases. A release is finalized, the semiconductor companies go to work developing chipsets, and the ecosystem needs to develop to produce networking and user equipment, software, and applications the time elapsed between the finalization of a release and the actual availability of products is typically a year and a half, at minimum. The 5G standard is currently in the first phase, Release 15, which is focused on eMBB. Service levels like URLLC and mMTC won’t be available until Release 16 (finalized July 2020; estimated chipset availability 2022/2023), Release 17 (scheduled to be finalized December 2021; device availability 2023/2024), and Release 18 (still in process). New capabilities will be added with each release, chipsets and user equipment will need to be upgraded while network equipment will undergo regular updates to support the new features.
Finally, although 5G will eventually be able to deliver all three service levels discussed above, 5G NR design (and the laws of physics) mean it will not be able to deliver all three of them simultaneously. Instead, 5G takes advantage of network slicing to target a specific service level. A link can be very high speed, for example, but not ultra-low latency. It might support mMTC but not at 10 Gbps. “You're going to have to pick your battles to a certain degree, but that’s okay,” says Daeuble. “In terms of industrial applications, very often the things that need high data rates don't need low latency but-but the machines that need low data rates are the ones that need very low latency as well. “
Achieving the performance enhancements and capabilities discussed above required completely revamping the over-the-air 5G interface. To ease the transition from 4G LTE to 5G and get at least some incremental performance improvements on the market quickly, the 3GPP established two deployment modes: non-standalone (NSA) and standalone (SA). In NSA mode, mobile operators use the existing 4G LTE network infrastructure as the anchor for control-plane tasks, while the 5G network basically functions as a data pipe to serve the user plane. In SA mode, the 5G network provides not just data services but the infrastructure for a major performance boost. “You’ll get significantly lower latency compared to the 5G NSA network,” says Harald Remmert, senior director of technology, Digi International (Hopkins, Minnesota). “In today’s cellular networks and 5G NSA, you might see 20 ms to 30 ms latencies. That’s already great for the public cellular network but it’s too much for an industrial controls network. When you look at 5G SA, you will eventually get down into latencies at the single-digit millisecond level.”
5G NR as defined by Release 15 currently only supports NSA mode, which is why the focus is solely on high-speed operation. The modifications necessary for SA mode, and the service levels it supports, will not be rolled out until Releases 16-18.
Any discussion of 5G NR needs to start with spectrum. Traditional licensed spectrum used for cellular (low band, 1 GHz and below) tended to be allocated in narrow channel widths, rather than the 100-MHz plus channel widths supported by 5G. In order to increase data rate, the 5G specification not only includes unlicensed spectrum in the 5 GHz and 6 GHz bands but extends out to the centimeter- and millimeter-wave regions (see table 4). The higher the frequency, the faster the data rate, although the trade-off is that higher frequencies are less effective at penetrating obstructions. Low-band cellular signals can pass through multiple walls, for example, while millimeter-wave signals require line-of-sight transmission.
Like 4G LTE Advanced and later, 5G can operate in anchored NR-U mode, using licensed spectrum for the anchor carriers and unlicensed spectrum for user-plane transmissions. Those anchor carriers can either leverage the LTE core in NSA deployments or use the 5G core in SA deployments. The LBT mandate remains in place to help prevent any existing Wi-Fi networks from being stepped on. Operators also have the option of standalone NR-U, in which only unlicensed spectrum is used for both control-plane tasks and user-plane transmissions. Note: Unlicensed spectrum capabilities will not be available until Releases 16.
The 6-GHz band is of particular interest for IIoT applications and other types of private industrial networks. Some industrial automation use cases like dockside cranes are outdoors. Many more are inside, where barriers abound ranging from walls and columns to massive pieces of machinery. The environment also tends to be dynamic, whether that involves warehouse shelves that are empty one day in full the next, forklifts, or two-ton pieces of structural steel being carried down the production floor by overhead crane. Although the millimeter-wave band offers capacity and speed, its inability to penetrate objects will make it more challenging for changing environments.
Ultimately, the ideal industrial network will need to operate over a combination of frequencies. “You can have a base layer of connectivity in the low band, so a sub- gigahertz cellular and then capacity in the mid-band from 2 GHz to 6 GHz,” says Remmert. “Then, when you need even higher speed or more capacity, you can go into the millimeter-wave band. We believe that gives you the best resiliency.”
Because of the differences in capabilities and use cases for the different spectrum bands, the 5G standard defines two frequency ranges: FR1, which encompasses low-band and mid-band; and FR2, which encompasses the high band. Each band is treated differently in terms of over-the-air interface and channel parameters (see table 5).
For modulation, 5G NR leverages 256-QAM for both uplink and downlink. It applies carrier aggregation, but with much more flexibility. 5G NR supports channel bandwidths of up to 400 MHz, depending on the frequency range, and increases the number of component channels that can be aggregated from five to eight. As a result, a single aggregated carrier can combine up to 3.2 GHz of bandwidth. Channels can be aggregated from any frequency bands within a given frequency range (FR1 or FR2). NSA network deployments can even aggregate 4G and 5G component carriers together.
5G NR uses cyclic prefix OFDM (CP-OFDM) for downlink and uplink. Unlike 4G LTE, which fixes subcarrier spacing at 15 kHz, 5G NR supports scalable subcarrier spacing Δf
where numerology n=0, 1, 2, 3, 4. This change has many ramifications for the performance of the network. Δf is the reciprocal of symbol transmission time interval (TTI) – doubling the subcarrier spacing, for example, cuts TTI in half. This is a key mechanism behind tuning the network to deliver eMBB or URLLC.
Like 4G LTE, 5G defines 10-ms frames that are each divided into 1-ms sub-frames. 4G LTE defined one slot per subframe. In 5G, each subframe is divided into a variable number of timeslots that each includes 14 symbols. Increasing subcarrier spacing reduces TTI, which reduces slot length and the duration of the subframe, reducing latency. This is an example of one of the ways that scaling subcarrier 5G and are enables performance to be tuned for the different service levels. URLLC capabilities will first be available with Release 16 and will truly come into their own with Release 17.
To address the skyrocketing growth of the IIoT, 5G NR applies massive MIMO (mMIMO). In this version, multiple antennas on the base station communicate with multiple UEs, expanding coverage. For FR1, the standard specifies 8x8 MIMO, which corresponds to 64 antennas at the base station. Although it can be operated in multiple modes, including SU-MIMO, industrial applications will get the biggest bang for the buck from operating in multi-user mode. In this case, each antenna targets a different device. The spatial diversity (multiple paths for the same data stream) can be used to minimize interference.
5G also applies another technology for MU-MIMO: beamforming (see figure 4). Instead of broadcasting to a large area incorporating multiple UEs, each antenna broadcasts a targeted beam aimed at a single QE or small collection of UEs. Beamforming helps improve transmission, even in environments with obstructions and reflective surfaces. It is particularly important for millimeter-wave applications.
5G was designed from the beginning to be a reconfigurable network. It includes an abstraction layer above the hardware layer that enables 5G to support software-defined radio (SDR). Coupled with some of the flexible characteristics of 5G NR, this makes it possible for network operators to define network slices that provide specific service levels. One slice might be configured to deliver eMBB performance, while another would be set up for URLLC.
This article reviewed a few of the more important technical innovations that power 4G LTE advanced Pro and 5G performance. 5G is still in its early stages but when fully developed shows every sign of living up to the hype. “4G was defined for consumers for mobile broadband,” says Peter Linder, head of 5G marketing at Ericsson (Plano, Texas). “5G was designed from the beginning for business and government use. It’s designed to support fixed wireless broadband as well as mobile. Further networks will evolve from just supporting universal data access, i.e. One slice fits all, to also support business- and mission-critcal connecivity where latency, reliability and availability is key.”
4G LTE should not be dismissed too quickly, however, cautions Daeuble. "From R11-12 all the way to R15, 3GPP has added a lot of features to LTE for critical communications and industrial needs, many of which have paved the way for 5G improvements,” he says, pointing to LTE-M+NB-IoT (low power sensors), reliability features, improvements to latency, etc. “Used as part of a private wireless network, 4G LTE technologies can provide significant improvements in coverage, performance, reliability, security, and mobility vs other wireless technologies.”
To learn more about when to consider 4G and 5G, and how to successfully build out a private cellular network, see part three of this series, to come in July 2021.
Thanks go to Justin Shade, senior product marketing specialist at Phoenix Contact, for useful conversations.