Chapter 8
Cellular Networks for IoT
Every new generation of wireless networks delivers faster speeds and more func- tionality through our smart phones. 1G brought us the very first cell phones, 2G let us text for the first time, 3G brought us online and 4G delivered the speeds that we enjoy today. But as more users come online 4G networks have just about reached the limit of what they’re capable of at a time when users want even more data for their smart phones and devices. Now we are headed towards 5G, the next generation of wireless. In this chapter, we will review the evolutions in the mobile cellular standards and shed some lights on the recent advancements toward enabling massive and critical IoT application in cellular systems.
8.1 The Cellular Concept
Wireless cellular system has been considered as the most promising solutions for emerging IoT applications and services. Cellular systems have evolved sig- nificantly since the launch of the 1st generation of mobile standard in 1980s.
Cellular systems are operated over licensed spectrum, that in reserved for only mobile technologies. In cellular systems, the area that should be services is splitted into cells. A base station is installed at the centre of each cell and each cells is assigned with multiple frequencies. Providing that adjacent cells are not using the same frequency, the inter-cell interference can be minimized. However, cells that are far from each other can re-user the same set of frequencies. This is called frequency reuse which is the main enabler of the cellular concept.
Each of these cells is assigned with multiple frequencies which have corre- sponding radio base stations. The group of frequencies can be reused in other cells, provided that the same frequencies are not reused in adjacent neighbor- ing cells as that would cause co-channel interference. Unfortunately, there is inevitably some level of interference from the signal from the other cells which use the same frequency. This means that, in a standard FDMA system, there must be at least a one cell gap between cells which reuse the same frequency.
115
116 CHAPTER 8. CELLULAR NETWORKS FOR IOT
Frequency reuse is mainly enabled by the propagation loss in wireless chan-
nel. That is the power of RF signals significantly reduce when travelling long
distances. In practice, the signal power is proportional to 1 , where d the dis- dα
tance between the transmitter and receiver and α is the path loss exponent. In practice alpha ranges between 2 and 5. This means that the power of the signal transmitted by the base station or a user at cell i is received by the cell j which is far from the cell i is negligible. Therefore, the same frequencies as in cell i can be reused in cell j, since the interference level is very low.
Frequency reuse enabled cellular systems to accommodate a large number of users at the same time over the same frequency bands providing that they are far from each other. It is clear that users that are in the same cell cannot use the same frequency as their transmission will interfere with each other. Users in the same cell share the channels using TDMA, FDMA, or CDMA, or OFDMA.
Different cells are connected (usually via fibre link for low latency communi- cations) to a central server to coordinate the transmissions and frequency and resource allocations. This server coordinate the mobile handover in a way that the user experiences a seamless connection to the server. Mobile handover takes place when a mobile user movers from one cell to another or the server decides to offload a certain base station. Fig. 8.1 shows the basic concepts in cellular systems.
Figure 8.1: Wireless cellular systems enabling concepts.
8.2 The 1st Generation
The first commercially cellular network, the 1G generation, was launched in Japan by Nippon Telegraph and Telephone (NTT) in 1979, initially in the metropolitan area of Tokyo. Within five years, the NTT network had been expanded to cover the whole population of Japan and became the first nation- wide 1G network. 1G is an analog systems which users FDMA to accommodate multiple users in a cell. Fig. 8.2 shows how the available bandwidth is divided
8.3. THE 2ND GENERATION 117
and allocated to different users. The cell capacity, defined as the number of users that can be simultaneously serviced in a cell, in 1G is limited as a large frequency gap should be considered between the subbands. Each subband could only carry one analog voice call. 1G service was unable to carry data.
Another factor that limited 1G is the analog devices. Analog devices in general are heavy, power inefficient, and usually expensive, therefore 1G systems could not scale well.
Figure 8.2: The 1st generation of mobile standard.
8.3 The 2nd Generation
The digital transformation revolutionized the cellular systems. The 2nd gener- ation of mobile standards uses digital signal rather than analog signals. Digital circuits are usually weight less and cost much less compared to their analog counterparts. 2G systems is more scalable compared to 1G and mobile hand- sets became much smaller in size. Digital signals can be also protected much more easily compared to analog signals. In 2G the voice signal is encoded using the voice encoder to convert analog signal to the digital signal. Using appro- priate sampling the voice signal can be compressed, therefore less bandwidth is required to transmit the voice without a noticeable sacrifice in quality. See Fig. 8.3.
2G uses the same bandwidth for each subband (i.e., 30kHz), but allowed for time division multiple access. That is the same frequency can be used in the same cell by more than one user when they are separated in time. This significantly improved the cell capacity. The first release of 2G allowed three users per radio channel of bandwidth 30kHz in each cell using TDMA. Later, by introducing GSM higher bandwidth (i.e., 200kHz) was considered for each radio channel and 8 users were allowed to share that channel via TDMA. Huge im-
118 CHAPTER 8. CELLULAR NETWORKS FOR IOT
Figure 8.3: The 2nd generation of mobile standard.
provement in the cell capacity was achieved by introducing GSM. Data service was introduced for the first time in mobile services in 2G. GPRS allowed for simple data service, including text and control messaging. Fig. 8.4 shows the advancements in the 2G era. In 2G the voice signal is encoded using the voice
Figure 8.4: The 2nd generation of mobile standard.
encoder to convert analog signal to the digital signal. Using appropriate sam- pling the voice signal can be compressed, therefore less bandwidth is required to transmit the voice without a noticeable sacrifice in quality. See Fig. 8.3. 2G is based on TDMA in the cells, that is a large gap between subbands in each cell is required. That limits the scalability of 2G systems.
8.4. THE 3RD GENERATION 119 8.4 The 3rd Generation
The third generation of mobile standard enables the emergence of smart phones which provided seamless and simultaneous voice and data connectivity for a large number of users. At the same time, consumers were provided with broad- band internet access in offices and home. Mobile users demanded more data over their phone. Significant advancements in device technology resulted in the era of smart phones.
The latest release of 2G enabled simple text messaging, email exchange, and news headlines. That was enabled by having some of the optimized voice channels reserved for data transmission. CDMA was also introduced in the latest versions of 2G which increased the cell capacity.
3G introduced an optimized data channel for data only. Users therefore were provided huge bandwidth for data transmission. This enabled multimedia transmission, web browsing, Apps and navigation systems on handheld devices (See Fig. 8.5).
Figure 8.5: The 3rd generation of mobile standard.
The latest release of 3G increase the bandwidth to 5MHz to even further speed in data transmission. Using the wide band CDMA and adaptive resource allocation, those channels reserved for voice transmission could be used for data transmission whenever they are unoccupied. This increased significant improve- ment in data rate and enabled streaming. Fig. 8.6 compares different release of 3G.
8.5 The 4th Generation
In 4G the bandwidth was further increased to 20MHz and the use of orthogonal FDMA (OFDMA) enabled higher data rates. Using OFDMA and carrier ag-
120 CHAPTER 8. CELLULAR NETWORKS FOR IOT
Figure 8.6: The 3rd generation of mobile standard and WCDMA.
gregation, different radio channels can be combined and allocated for a user for data transmission. This flexibility in 4G enabled high speed data transmission suitable for video streaming and other data hungry applications.
Figure 8.7: The 4rd generation of mobile standard.
As can be seen in Fig. 8.7 using advanced multiple antenna technologies was another enabler of the 4G standard. Multiple antennas at the base stations and users enabled spatial diversity and multiplexing gain, which significantly improved the quality of service in 4G systems. In other words, 4G improvements in data rate in multiple folds. Let us have a look at the Shannon Capacity formula in (8.1)
C = W × n × log2(1 + SINR), (8.1)
where W is the bandwidth, n is the number of antennas, and SINR is the signal to noise plus interference. 4G increased the bandwidth, increased the
8.6. THE 5TH GENERATION 121
number of antennas, and using advanced signal processing and multiple antenna technologies, significantly increased SINR. Fig. 8.8 summarizes the evolution of mobile standards.
Figure 8.8: The evolution of mobile standard.
8.6 The 5th Generation
The fifth generation of mobile standards have a fundamentally different focus compared to the previous generations. Up to 5G, the focus was to improve the capacity of mobile broadband. That is an advanced device, such as a smart phone controlled by human, is communicating with the base station and de- mands for voice and data service. 5G is still trying to increase the data rate for mobile broadband to enable online and realtime gaming, realtime 3D video streaming, and virtual reality applications. Beside this 5G introduced two other major service categories which have completely different requirements that have not been considered before. The first service category is massive IoT which de- mands for low power consumption, long battery life time, massive number of devices, low cost and flexibility. Another category is Critical IoT which de- mands for high reliability, availability, low latency, and in some cases high data rates. Fig. 8.9 shows different service categories and their requirements cur-
rently considered in 5G. This video u also sheds some light in the 5G vision.
As can be seen in Fig. 8.9 in the 5G vision, the enhance mobile broadband targets multi Gbps data rates, extreme capacities to accommodate many users and therefore very high data rates per square kilometer, and deep awareness. On the other hand, massive IoT targets for ultra low energy consumption, i.e., more than 10 years battery lifetime, ultra low complexity, ultra high density, i.e., 1 million devices per square kilometer, and deep coverage to provide deep indoor connectivity. Mission critical IoT demands for high security, ultra high reliability, ultra low latency, and extreme user mobility support. 5G promises
122 CHAPTER 8. CELLULAR NETWORKS FOR IOT
Figure 8.9: Different service categories in the context of 5G.
to provides connectivity for a diverse range of applications and services. See Fig. 8.10 for a summary of these applications.
Figure 8.10: 5G target applications and services.
8.6.1 Enablers of 5G
5G promises to provide multi-Gbps data rates for mobile broadband. This is not possible with existing infrastructure. A paradigm shift from current systems and protocols should happen to enable such a massive growth in capacity and services.
8.6. THE 5TH GENERATION 123
5G faces several challenges which have to be addressed simultaneously. The first challenge is that the number of users and devices have significantly increased compared to the previous generations of mobile standards. One reason for such growth in the number of users/devices is the increase in popularity of IoT services and applications. There are many more users now that user IoT services and each service introduce a large number of machine/devices to the systems. 4G systems are on the limits and cannot support more users/devices.
Another challenge is that the users still demand for higher data rates and that was further exacerbated by data hungry applications such as online gaming, virtual reality, etc. Current 4G systems simply cannot support such high data rates. 5G aims to accommodate 1000 times more users/devices and provide 10 times faster data rates compared to the latest 4G standard. With that speed, you can download an HD movie in under a second. 5G will be the foundation for virtual reality, autonomous driving, the Internet of Things, and stuff we can’t even yet imagine.
There are five brand new technologies emerging as a foundation of 5G. These are 1) millimeter waves, 2) small cells, 3) massive MIMO, 4) beamforming, and 4) full duplex. In what follows, we briefly explain each technology. This video
u summarizes the enabling technologies for 5G. Millimeter Waves
Your smart phone and other electronic devices in your home use very specific frequencies on the radio frequency spectrum typically those under six giga- hertz. But these frequencies are starting to get more crowded. Carriers can only squeeze so many bits of data on the same amount of radio frequency spec- trum as more devices come online. We’re going to start to see slower service and more draught connections. The solution is to open up some new real estate so researchers are experimenting with broadcasting on shorter millimeter waves those that fall between 30 and 300 gigahertz. This section of spectrum has never been used before for mobile devices and opening it up means more bandwidth for everyone but there is a catch. Millimetre waves can’t travel well through buildings or other obstacles and they tend to be absorbed by plants and rain. To get around this problem we’ll need technology number two.
Small Cell
Today’s wireless networks rely on large high-powered cell towers to broadcast their signals over long distances. But higher-frequency millimeter waves have a harder time traveling through obstacles which means if you move behind one you lose your signal. Small cells would solve that problem using thousands of low- power mini base stations. These base stations would be much closer together than traditional towers forming a sort of relay team to transmit signals around obstacles. This would be especially useful in cities as the user moved behind an obstacle his smart phone would automatically switch to a new base station in
124 CHAPTER 8. CELLULAR NETWORKS FOR IOT better range of his device allowing him to keep his connection.
Massive MIMO
Massive MIMO, stands for multiple-input multiple-output. Today’s 4G base stations have about a dozen ports for antennas that handle all cellular traffic but massive MIMO base stations can support about a hundred ports. This could increase the capacity of today’s networks by a factor of 22 or more. Of course massive MIMO comes with its own complications. Today’s cellular antennas broadcast information in every direction at once and all of those crossing signals could cause serious interference which brings us to technology number 4.
Beamforming
Beamforming is like a traffic signalling system for cellular signals instead of broadcasting in every direction. It would allow a base station to send a focused stream of data to a specific user. This precision prevents interference and it’s way more efficient that means stations could handle more incoming and outgo- ing data streams at once. Consider you are in a cluster of buildings and you are trying to make a phone call. Your signal is ricocheting off of surrounding buildings and criss-crossing with other signals from users in the area. A massive MIMO base station receives all of these signals and keeps track of the timing and the direction of their arrival. It then uses signal processing algorithms to triangulate exactly where each signal is coming from and plots the best trans- mission route back through the air to each phone. Sometimes it will even bounce individual packets of data in different directions off of buildings or other objects to keep signals from interfering with each other. The result is a coherent data stream sent only to you.
Full Duplex
Today’s wireless systems are based on Half Duplex, that is you cannot simul- taneously transmit and receive. A basic antenna can only do one job at a time either transmit or receive. This is because of a principle called reciprocity which is the tendency for radio waves to travel both forward and backward along the same frequency. To understand this it helps to think of a wave like a train loaded up with data. The frequency it’s traveling on is like the train track and if there’s a second train trying to go in the opposite direction on the same track you’re going to get some interference. Up until now the solution has been to have the trains take turns or to put all the trains on different tracks or frequen- cies. But you can make things a lot more efficient by working around reciprocity. Researchers have used silicon transistors to create high speed switches that halt the backward role of these waves. It is kind of like a signalling system that can momentarily reroute to train so that they can get past each other that means there’s a lot more getting done on each track a whole lot faster.
5G is still a work in progress and it will likely include other new technologies too and making all of these systems work together will be a whole other challenge
8.7. SPECIFIC SOLUTIONS FOR MASSIVE IOT 125 but if experts can figure that out ultra fast 5g service could arrive in the next
five years.
8.7 Specific Solutions for Massive IoT
Narrowband IoT (NB-IoT) is a new mobile network especially designed for the Internet of Things. it runs in licensed spectrum and is based on an international standard of 3GPP. With today’s networks as ITU, the data rates increase more and more up to a few hundred megabits per second, with narrowband IoT it’s just the opposite way around. NB-IoT supports a few kilobits per second. If you think about IoT devices as small and simple sensors isn’t that enough?
For example a water meter reports infrequently its counter only requiring a few bytes per day. On the other hand NB-IoT has some really impressive advan- tages. It provides a better coverage with deep indoor penetration, technically we talk about 20 dB plus on the link budget compared to GPRS. Furthermore a longer battery lifetime can be achieved up to 10 years with the equivalent of two AAA batteries and finally the customer below NB-IoT communication modules are expected to be cheaper than GSM modules and subscription fees will be lower than today.
NB-IoT can be used for smart parking systems, which can reduce traffic in the cities by communicating free on street parking lots. Smart waste manage- ment can have waste management companies to be more efficient and to reduce costs by measuring the fill outs. Smart metering reduces costs and improves the service quality by remotely reading out the devices.
NB-IoT is a communication standard designed for IoT systems operates over carrier networks, either in the existing GSM carrier network, or in the unused guard band in the LTE channel, or operates independently (See Fig. 8.11 for these implementation scenarios). The technology is currently part of the LTE specification.
Figure 8.11: NB-IoT implementation scenarios.
Another technology is LTE-M that is also a low bandwidth protocol for mas-
126 CHAPTER 8. CELLULAR NETWORKS FOR IOT
sive IoT. Nb-IoT and LTE-M are set to roll out in 2017 both are optimized for the needs of Internet of Things applications such as industrial controls, residen- tial security, smart metering, municipal infrastructure or precision agriculture. NB-IoT offers low bandwidth data connections at low cost and is currently Europe focused while LTE-M is optimized for a higher bandwidth and mobile connections including voice. It will start rolling out in North America. Table 8.1 shows the specification of LTE-M and NB-IoT.
Table 8.1: Comparison betweeb LTE-M and NB-IoT.
LTE-M
384 kbps
50-100 ms
Best at medium data rates Yes
Yes
1
North America
Peak data rate Latency Power consumption Mobility Voice Antenna Initial regions
NB-IoT
<100 kbps
1.5-10 seconds
Best at very low data rates No
No
1
Europe
Both LTE-M and NB-IoT introduced enhanced power save modes to improve the battery life time. Extended sleep cycles where considered in these technolo- gies as they are mostly targeting low traffic massive applications. Using low complexity and low power technologies they have also improved the coverage, so deep indoor communications are supported by them.
8.8 Further Reading
• 5G unlocks a world of opportunities: Top ten 5G use cases, Huawei, n
• The value of 5G for cities and communities, Telefonica UK, n.
• 5G radio access, Ericsson, n. M. R. Palattella et al., ”Internet of Things in the 5G Era: Enablers, Architecture, and Business Models,” in IEEE Journal on Selected Areas in Communications, vol. 34, no. 3, pp. 510-527, March 2016 n.
• Making 5G NR a reality, Qualcomm, n.
• Ian F. Akylidiz, et. al., 5G roadmap: 10 key enabling technologies, Com- puter Networks, vol. 106, pp. 17-48, Sep. 2016 n.