ECE-60022

Wireless Communications Networks

An Introduction

Professor Xiaojun Lin

January 13, 2020

2G/3G/4G/5G?
 Overall the past 20 years, wireless networks have
seen dramatic changes over several generations of
technology

 In comparison, the basic paradigm of the Internet

(i.e., TCP/IP) remains the same since 70s/80s

 What are the main drivers for these changes? What

are the main technologies in each generation of
wireless networks? What lessons have we learned?
 “Network” vs. Point-to-Point
Communications
 Prior to wireless networks, successful systems have been built
that utilize the capability of send information over wireless
signals
 Radio, Television, etc.

 AM (amplitude modulation), FM (frequency modulation),

phase modulation, etc.

 Utilize PHY layer that is point-to-point or point to multi-point

 “Network” is inherently different

 It must be able to support multiple point-to-point connections

simultaneously
 Multiple Access

 How many users should be allowed to transmit simultaneously

on a single unit of resource?
 How to handle the interference between then?
 How to allocate resources across users to meet their possibly
diverse application requirements?
 What is the “capacity” of the network as a whole?

As we review the advance of wireless networks, let us pay close

attention to how each generation of wireless networks address
these questions

There is perhaps one concept that represent the importance of

multiple access most prominently, i.e., cellular
 First Mobile Telephone Service (0G?)

 Introduced in the U.S. by AT&T (1946)

 Used to interconnect Mobile Users (usually in automobiles) to the
public telephone land-line networks.
 Systems used a single powerful transmitter from the base-station
to cover up to approx. 50 miles radius.
 Systems were based on Frequency Modulation (FM)
transmission.
 Each frequency channel used 120KHz of spectrum to transmit a
voice connection with an effective bandwidth of only 3KHz
 First Mobile Telephone Service (cont’d)

 Advanced systems for their time but clearly very inefficient.

 The first system can accommodate only 40 users at any time.

 In the meantime, demand for mobile telephone service grew

very quickly and stayed ahead of available capacity in many
large cities.
 Offered traffic load >> available capacity
 Hence, service was terrible (blocking probabilities were as high
as 65%, or even higher!)
⇒ Usefulness of mobile telephone decreased as users found
that blocking often prevented them from getting a circuit during
peak periods.
First Mobile Telephone Service (cont’d)
Telephone companies soon realized that for true mobile services to
develop:

 Large blocks of spectrum would be needed to satisfy the demand

of users, especially in urban areas.
 But the amount of usable spectrum is limited!

 Significant improvements need to be made in the usage of the

available spectrum.
Advent of Cellular Systems
 In the mid-1960s, Bell Systems introduced the
Improved Mobile Telephone Service (IMTS) with
 enhanced features such as: Direct Dialing, Automatic
Trunking, Full-duplex service
 a reduction in FM channel BW from 120Khz to 25-30KHz.

 However, the capacity problem is not fundamental

solved.

 In the late 1960s and early 1970s, work began on the

first cellular telephone systems.
Pre-Cellular Days
 As discussed earlier, the traditional approach to mobile radio
involved setting up a high-power transmitter located on top of
the highest point in the coverage area.

 This allows line-of-sight transmission to a large distance (as

much as 40-50 miles away for a very high base-station antenna).
 Similar to TV

 However, the result was that few available radio channels were
locked up over a large area
 Cellular Systems and Frequency Reuse
 The cellular concept handles the
coverage problem differently
 Use a large number of low
power transmitters, each
designed to serve only a small
area called a cell.
 Frequencies were not reused in
adjacent cells to avoid
interference
 However, the wireless spectrum
is reused among those cells that
are further away.

(Source: Prof. Vladimir V. Riabov)

 Cellular Systems and Frequency Reuse
 We will call a unit of wireless
spectrum a channel.
 Depending on the specific
system, a channel could be a
frequency (FDMA), time-slot
(TDMA), or a code (CDMA). The
concept of a channel allows us
to understand these systems
through a unifying framework.
Frequency Reuse Example

1
1 4 2
1 3 1
4 2 4 2 4 2
3 1 3
3 4 2
3

(Source: Prof. Vladimir V. Riabov)

Cell Splitting

(Source: Prof. Vladimir V. Riabov)

Can cells be split indefinitely?
 As cells become smaller, their shapes and locations become more
irregular (even more true for 4G/5G systems).

 It is also more difficult to place the base-stations and connect them to

the wired backbone

 Further, for voice calls, the problem now was that not all mobile calls
would be completed within a cell.

 When a mobile approached the boundary of its current cell, heading to

a new cell, some mechanism was needed in place to allow the mobile to
successfully end its communication with the base-station in its current
cell, and begin its communication with the base-station of its new cell.
This procedure is called handoff.
 Cellular: Summary
So, in summary, cellular is way to organize multiple
access so that channels can reused among cells to
improve network capacity.

However, cell splitting will lead to increased complexity

(handoff), and will eventually face practical limits.

In addition to adding spectrum and splitting cells, other

innovations in managing multiple access are needed to
meet the continuing demand for more network capacity.
Question
 So, what are the key distinguishing features of a
wireless system versus a wireline system (e.g., those
studied in ECE 547)?
Generations of Mobile Wireless Networks

1981 1992 2001 2011 2020

Analog voice Digital voice Data service OFDM mmWave
Cellular TDMA, CDMA Opportunistic MIMO Massive MIMO
FDMA scheduling HetNets MTC
NOMA

(Source: [Link])
First Generation Cellular Systems
 First generation cellular systems started appearing in
the 1970s
 Simultaneously in developed countries:
 Advanced Mobile Phone System(AMPS) in North America
 Total Access Communications Systems (TACS) in U.K.
 Nordic Mobile Telephone (NMT) in Scandanavian countries
 C450 in West Germany
 Nippon Telephone and Telegraph(NTT) in Japan
 All use analog voice and FDMA
Second Generation Cellular Systems
The principal goals of second-generation wireless
systems were:

 A digital system with higher capacity (hence lower

cost) derived from digital techniques including
efficient speech coding, error-correcting channel
codes, and bandwidth-efficient modulation
techniques.

 To offer, in addition to voice services, some

rudimentary data services.
 Technologies for Second-Generation
Cellular Systems
Several standards and system design were proposed for
competing technologies for second-generation cellular
systems.
 FDMA (N-AMPS): North America

 TDMA/FDMA (GSM): Europe (where the need for a

common standard is higher)

 CDMA spread spectrum (IS-95): North America

Mostly two standards prevailed: GSM (a TDMA/FDMA)

standard and IS-95 (a CDMA standard)
 GSM (Example of a TDMA/FDMA system)
 GSM (Global System for Mobile Communications) was driven by the
need for a common mobile standard throughout Europe and the desire
for digital transmission compatible with data and privacy.

 Spectrum was reallocated for much of Europe near 900MHz so that

completely new technology could be developed around GSM.

 The GSM effort in the mid-1980s considered several system

implementations including TDMA, CDMA, and FDMA technologies. In the
end, a TDMA/FDMA technology was chosen with a radio link-rate of
270kbs.

 The GSM standard is important and many third and fourth generation
standards will be extensions of it.
FDMA/TDMA in GSM

A channel

1 200Khz
4 2 per band
3

25M/200k -> 125 FDMA bands

8 TDMA channels per FDMA band

(Source: Prof. Vladimir V. Riabov

and [Link])
CDMA in IS-95

 All frequency channels are

reused in all cells
1  The frequency bands are
1 1
1 1
also wider (1.25Mhz), so
1 1 1 multiple users transmit in
1 1 1 the same band
1 1
1 1 simultaneously
1  Interference between
users is suppressed by a
technique called direct-
sequence spread-spectrum
(DSSS)
CDMA in IS-95

 The number of users that

can simultaneously
1 transmit now depends on
1 1 how much interference is
1 1
1 1 1 suppressed by DSSS
1 1 1
1 1
1 1
1
 Code Division Multiple Access (CDMA)

 The claim by Vitterbi (CEO of QUALCOMM) was that,

because CDMA could use frequencies in all cells
(universal reuse), the capacity improvements in
CDMA were almost 30 times of TDMA/FDMA systems.
 We will examine the CDMA capacity later in the class

 CDMA also claims better multipath resistance,

superior voice quality, increased call privacy, and soft
handoff.

 This frenzy about CDMA continues into 3G when all

3G standards are based on CDMA, but …
Generations of Mobile Wireless Networks

1981 1992 2001 2011 2020

Anolog voice Digital voice Data service OFDM mmWave
Celluar TDMA, CDMA Opportunistic MIMO Massive MIMO
FDMA scheduling HetNets MTC
NOMA
 Third Generation Wireless Systems
 First-and second-generation digital systems are designed
to support voice communication with limited data
communication capability.

 Third generation systems (and beyond) are targeted to

offer high-data-rate communications and expect to
support a wide variety of applications

 What was the killer app for 3G?

Third Generation Wireless Technologies
The development of 3G is carried out by the
standardization procedure called IMT-2000
(International Mobile Communications 2000) initiative.

The standardization procedure was done in parallel in

different parts of the world and the eventual standards
are mainly:
 Extensions of GSM -2.5 G

 Wideband CDMA systems -3 G.

 Extensions of GSM (2.5G): GPRS/EDGE

 Generalized Packet Radio Service (GPRS) is an enhancement of GSM

 GPRS provides packet-switched data services at 5K-100Kbps by
reusing the existing narrow-band GSM channels/infrastructure for
circuit-switched services such as voice telephony.

 Enhanced Data Rates for GSM Evolution (EDGE) uses adaptive

modulation techniques and can produce 384Kbps maximum
transmission.
 Both GPRS and EDGE aim not to replace the current infrastructure,
but to co-exist with it.

 To achieve Mbps bit rates will require using more sophisticated

physical layer as well as resource & radio management.
Wideband CDMA
In March 1999, the final decision was made to select wideband
CDMA as the access scheme for 3G, with three optional modes:

 W-CDMA (or UMTS): CDMA Direct Spread. 5MHz channel.

Frequency Division Duplex(FDD).Based on Europe (ETSI) and
Japan (ARIB)’s FDD proposal.

 CDMA-2000: CDMA Multi-carrier. Can use multiple 1.25Mhz

channels. Based on US (TIA)’s proposal.

 TD-SCDMA:CDMA TDD (time division duplex). Can use multiple

time slots. Based on China’s proposal.
Wideband CDMA
 W-CDMA is meant to serve as an evolution path for GSM (GSM
->GPRS/EDGE->W-CDMA). It uses much of same core network
as GSM.

 CDMA 2000 is meant to serve as an evolution path for IS-95

(including 1xRTT, EV-DO).

 Downlink speed can reach 10+Mbps.

 Key Technology in 3G Systems
 However, under the disguise of CDMA is a very different way of
resource/radio management and spatial reuse

 In each cell and channel, at most one user is active at a time

 CDMA is only to suppress interference from other cells
 (We will study the reasons behind this change of direction)

 The single user will transmit at a higher rate if its channel condition is
good (e.g., closer to the base-station) and will get higher data rate
 Adaptive modulation and coding (AMC)

 Among many users, the one with best channel condition will be
scheduled to transmit
 Opportunistic scheduling to exploit diversity in time
Generations of Mobile Wireless Networks

1981 1992 2001 2011 2020

Anolog voice Digital voice Data service OFDM mmWave
Celluar TDMA, CDMA Opportunistic MIMO Massive MIMO
FDMA scheduling HetNets MTC
NOMA
 4G Wireless Systems
 Even higher data-rate: envisioned to attain 100Mbps for
high mobility, 1Gbps for low mobility

 CDMA is abandoned. OFDM is used instead. (Will study

the reason why.)

 High spectral-efficiency (via OFDM, MIMO)

 All IP, packet switched

 4G Wireless Systems
 Two competing standards:
 LTE (Long Term Evolution) by 3GPP: driven by telecom industry.

 WiMax (802.16): driven by computer vendors.

 Both based on OFDM

 LTE prevails and becomes the first universal standard across the
globe

 With 100 MHz aggregated bandwidth, LTE-Advanced provides almost

3.3 Gbit peak download rates per sector of the base station under
ideal conditions. (Typically rate may be lower, using 20Mhz band.)

 Voice over LTE will eventually get rid of legacy circuit-switched

systems altogether.
Key Technology for 4G LTE
 OFDM: more efficient frequent-time modulation

 MIMO: Multi-Input Multi-Output communications

 HetNets: Heterogeneous networks of many small cells

 Orthogonal Frequency Division
Multiplexing (OFDM) in 4G LTE
 OFDM divides a wide band
0.5ms or 1 ms
into a number of narrow-
15khz band sub-carriers, each of
which carries information
 OFDMA allow each user to
be scheduled on a subset
of frequency-time slots.
 Opportunistic
scheduling can now
exploit diversity in both
frequency and time
Source: [Link]
 OFDMA in 4G LTE

OFDMA in 4G LTE: TDMA/FDMA in 2G GSM

- higher peak rate
- better handling of freq. selectivity
- higher spectrum efficiency
 Multi-Input Multi-Output (MIMO)
Transmissions

Source: wikipedia

 𝑁𝑡 antennas at the transmitter, 𝑁𝑟 antennas at the receiver

 If the channel matrix is of full rank, min 𝑁𝑡 , 𝑁𝑟 data
streams can be sent in parallel (known as spatial
multiplexing)
 Potential data rate increases linear with the antenna size
min 𝑁𝑡 , 𝑁𝑟
 Multi-Input Multi-Output (MIMO)
Transmissions

 MIMO is said to exploit spatial diversity. (In contrast,

opportunistic scheduling exploits frequency- and time-
diversity)
 Need to know the channel gain between every pair of
transmit-/receive-antennas (CSI): Significant overhead in 4G
 Multi-User (MU-) MIMO can benefit even if the receiver only
has a single antenna
 Heterogeneous Networks (HetNet)
 Smaller cells are needed to accommodate increasing traffic
demand: micro/pico/femto cells.
 The result is highly irregular topology and load
 Challenge: more like an ad hoc network. Traditional network
planning and control based on cellular no longer work! (See “Seven
ways that HetNets are a Cellular Paradigm Shift”)

Pico
Macro cells
cells

Femto
cells
(Source: Prof. Jeffery Andrews, UT Austin)

Traditional Cellular Networks Heterogeneous Cellular Networks

Generations of Mobile Wireless Networks

1981 1992 2001 2011 2020

Anolog voice Digital voice Data service OFDM mmWave
Celluar TDMA, CDMA Opportunistic MIMO Massive MIMO
FDMA scheduling HetNets MTC
NOMA
How about 5G?

 First specification (Phase 1) agreed on Dec. 2017

 Wish Lists (3 Use Cases) for 5G
 High Data Rate: Enhanced mobile broadband
(eMBB)
 Low Delay: Ultra-reliable low latency
communications (URLLC)
 Massive # of devices:
machine-type
communications (mMTC)
 Possibly along different
“slices” of the same
core network (network
virtualization)
Source: Samsung
High Data Rate

Source: Samsung

What driving applications?

Ultra Low Latency

Source: Samsung

What driving applications?

Ultra Low Latency
 Virtual Reality, Augmented Reality
 Automation, robotics, remote-control
 Massive Connectivity
 Internet-of-Things
 But each device may
be of low rate

 Low overhead is
critical
 Low latency
 Low power

Source: [Link]
 Some Likely Technology for 5G

Source: [Link]
 mmWave
 30-300Ghz, sometimes including 24Ghz (about 10mm and
below)
 Much higher free-space loss
 Absorption by atmosphere, rain, etc.
 Line-of-sight propagation: low diffraction, does not penetrate
walls/objects
Some Likely Technology for 5G

Source: [Link]

 mmWave
 Smaller coverage (100-200m), higher frequency reuse
 Highly-directional beams may be needed to improve
coverage
 Beam alignment and beam tracking becomes a problem
 Some Likely Technology for 5G

Source: Massive MIMO for 5G,

IEEE 5G Tech Focus: Volume 1,
Number 1, March 2017
 Massive MIMO
 Smaller wavelength of mmWave allows large antenna
arrays to be packed at the transmitter/receiver
 Potentially huge spectrum efficiency gain
 High directionality
 More resources needed for channel estimation/feedback.
Pilot contamination
Some Likely Technology for 5G
 Non-orthogonal access (NOMA)
 In 3G/4G each user is assigned orthogonal resource
(time/frequency) when transmitting
 Signaling is needed for scheduling resources
 For low-rate IoT applications, such signaling incurs high
overhead
 Potentially more advantageous to have multiple
transmissions over the same set of the orthogonal
resources

 Sounds like CDMA?

 How about spectrum efficiency, diversity gain, etc?
Generations of Mobile Wireless Networks

1981 1992 2001 2011 2020

Anolog voice Digital voice Data service OFDM mmWave
Celluar TDMA, CDMA Opportunistic MIMO Massive MIMO
FDMA scheduling HetNets MTC
NOMA
Technologies Competing with Cellular
Systems
 Wireless LAN
 Bluetooth
 Ad Hoc Networks

 Unlicensed band
 Very different multi-access technology: e.g., CSMA
 Often ad-hoc deployed (somewhat like HetNet now)
 Will discuss later in random access
 In Summary

 Demand for mobile network service keeps increasing

 Although the cellular concept prevails, a lot of
change in the underlying technology for multiple
access and resource management
 Why are particular changes are made in a new
standard?
 We need to understand the implication to multiple
access and how it impacts the system capacity
 Two High-Level Considerations

 Whether to design for the worst-case or for the

average-case?

 How to deal with variations in the channel quality

(i.e., diversity)?
 Many Detailed Questions
 How to maximize the network-layer benefits (e.g.,
capacity, delay) of various new PHY-layer technologies?

 How to manage highly irregular/ad-hoc topology and load?

 How to achieve various types of Quality-of-Service for

different applications (e.g., high throughput versus low
delay)?

 How to accomplish all these with low complexity and

overhead?
Overview of the Rest of the Class
 Understanding the wireless channel
 Cellular capacity and cellular channel allocation (Voice
systems)
 Optimal capacity and scheduling in broadband
wireless networks (Data systems)
 Stochastic geometry and scaling laws
 Case studies of 2G/3G/4G systems (with student
participation)
 Random access (for both wireless LAN and low-
overhead IoT access)
 Wireless control algorithms for low-delay
 Project presentation
Questions?
Learning Objectives
 Identify the key technology advances in each
generation of wireless networks
 Identify key considerations for multiple access and
resource management in wireless networks
 Identify the challenges ahead for 5G and Internet-of-
Things