ISLAMIC UNIVERSITY OF TECHNOLOGY

ORGANISATION OF ISLAMIC COOPERATION (OIC)

GAZIPUR, BANGLADESH

EEE 4541
Wireless Communication

Adaptation with Radio Link Quality

Dr. Mohammad Tawhid Kawser

Professor, EEE Dept.
Key Performance Indicator (KPI)
The common KPIs in communication links are:

– Throughput (Data Rate)

– Bit Error Rate (BER) /Bit Error Ratio (BER), Block
Error Rate (BLER), Packet Error Loss Rate (PELR)
– Latency (Delay in data transfer)
 Latency (Book: Peterson)
• The signal has propagation delay:
Propagation = Distance/Velocity of EM wave
delay

• Depending on the available capacity of the link, the data needs

some time to flow:
Transmit = Size/Bandwidth

• There can be Queuing delays because the packets may be

stored temporarily before they can be forwarded.

Latency = Propagation + Transmit + Queue

3
 Delay × Bandwidth Product (Book: Peterson)
The amount of data in transit (data existing on the link) is given by
Delay × Bandwidth Product
where Delay is the Propagation Delay.

The sender must transmit this amount of data before the first bit arrives at
the receiver.

A link with a latency of 50 ms and a bandwidth of 45 Mbps is able to hold:

50 ×10−3 sec × 45 × 106 bits/sec = 2.25 × 106 bits, or approximately 280 kB
of data.

4
Digital Modulation
Digital-to-analog modulation

Types of digital-to-analog modulation

 Aspects to digital-to Analog conversion

Baud/Symbol Rate
– Bit rate is the number of bits per second. Baud/Symbol rate is the number
of symbolsper second. Baud rate is less than or equal to the bit rate.
– Baud/Symbol rate is important in data transmission.
• Baud rate determines the bandwidth required to send signal
– Baud/Symbol rate = bit rate / # bits per symbol

– The bit rate of a signal is 3000. If each symbol carries 6 bits, what is the
baud rate?
• Baud rate = 3000/6 =500 bauds/sec or symbol/sec or sps
 Amplitude Shift Keying (ASK)

 A cos(2Πf c t ) binary1
s (t ) = 
 0 binary 0

On/Off keying (OOK)

• The strength of the carrier signal is varied to represent binary 1 and 0.

• Frequency and phase remains the same.
• Highly susceptible to noise interference.
 Frequency Shift Keying
• Frequency of the carrier is varied to represent digital data (binary 0/1)
• Peak amplitude and phase remain constant.
• Avoid noise interference by looking at frequencies (change of a signal) and
ignoring amplitudes.
• Limitations of FSK is the physical capabilities of the carrier.
• f1 and f2 equally offset by equal opposite amounts to the carrier freq.

 A cos(2Πf 1t ) binary1
s (t ) = 
 A cos(2Πf 2 t ) binary 0
Phase Shift Keying

• Phase of the carrier is varied to

represent digital data (binary 0 or 1)
• Amplitude and frequency remains
constant.
• If phase 0 deg to represent 0, 180
deg to represent 1. (2-PSK)
• PSK is not susceptible to noise
degradation that affects ASK or
bandwidth limitations of FSK
4-PSK (QPSK) method
 8-PSK

• We can extend, by varying the the signal by shifts of 45 deg (instead

of 90 deg in 4-PSK)
• With 8 = 23 different phases, each phase can represents 3 bits (tribit).
Quadrature Amplitude Modulation (QAM)

• PSK is limited by the ability of the equipment to distinguish between

small differences in phases.
– Limits the potential data rate.
• Quadrature amplitude modulation is a combination of ASK and PSK so
that a maximum contrast between each signal unit (bit, dibit, tribit,
and so on) is achieved.
– We can have x variations in phase and y variations of amplitude
– x • y possible variation (greater data rates)
• Numerous variations. (4-QAM, 8-QAM)

# of phase shifts > # of amplitude shifts

8-QAM and 16-QAM
 Bit Baud comparison (examples)

• A constellation diagram consists of eight equally spaced points on a circle. If

the bit rate is 4800 bps, what is the baud rate?

– The constellation indicates 8-PSK with the points 45 degrees apart. Since 23 = 8, 3
bits are transmitted with each signal unit. Therefore, the baud rate is 4800 / 3 =
1600 baud

• What is the bit rate for a 1000-baud 16-QAM signal.

– A 16-QAM signal has 4 bits per signal unit since log216 = 4. Thus, (1000)(4) = 4000
bps

• Compute the baud rate for a 72,000-bps 64-QAM signal.

– A 64-QAM signal has 6 bits per signal unit since log2 64 = 6.

– Therefore, 72000 / 6 = 12,000 baud
Impact of SNR on the constellation diagram
 Error Vector Magnitude (EVM)

An error vector is a
vector in the I-Q plane
between the ideal
constellation point and
the point received by the
receiver. In other words,
it is the difference
between actual received
symbols and ideal
symbols. Its magnitude
(calculated as RMS) is
Error Vector Magnitude
(EVM).
Symbol Error Probability vs. Eb/NO
Handling Packet Errors
 BER
– The Bit Error Ratio (BER) is the number of bit
errors divided by the total number of
transferred bits during a particular time
interval.
– The Bit Error Rate (BER) is the number of bit
errors per unit time.
 BLER
– Block Error Rate (BLER) is a ratio of the number
of erroneous blocks or packets of data to the
total number of blocks transmitted.
– Packet Error Loss Rate (PELR) is similarly
defined for packets of data.
 Error Detection in Packets
• Popular methods:
• Cyclic Redundancy Check (CRC) [used in Ethernet]
• Checksum [used in TCP]
• The techniques create small bit string which is
appended to the packet before transmission. The
receiver performs opposite calculation between
that field and the rest of the packet. If there is
mismatch, the packet is considered to include bit
errors.
 Error Correction in Packets
There are two fundamental methods:
1. Forward Error Correction (FEC) or using Error
Correction Code (ECC).
2. Retransmission of the erroneous packets (when the
packet errors are detected).

• Forward error correction (FEC) attempts to deliver

correct data packets but when it fails, the incorrectly
received packets need to be retransmitted.
• Retransmissions, in both uplink and downlink, can be
performed at different layers (e.g. TCP layer).
 Forward Error Correction (FEC)
• In FEC, the errors can not only be detected but also,
be corrected.
• Packets are coded before transmission, which
increases packet size. The additional length of the
packet is considered to be parity bits. The receiver
performs opposite type of calculation. If the bit
errors are not too much, they can be corrected.
 Examples of ECC codes
• Reed–Solomon code
• Hamming code
• Hadamard code
• Golay code
• Reed–Muller code
• BCH code
• Convolutional code (popular in 3G)
• Turbo code (popular in 4G)
• Low-density Parity-check (LDPC) code (popular in 5G)
 Explanation of ECC codes
• Repetition code is the
simplest ECC code.
• In this example, 3 bits
are sent with 2 parity
bits.
• If 5 bits were sent
(with 4 parity bits),
the error recovery
would be more
capable.
 Code rate of ECC codes
Code rate
= No. of actual information bits/No. of bits in the
packet

Lower code rate has higher reliability but larger packet

size.
 Receiver Sensitivity
It’s the minimum signal strength required at the receiver to get usable data without
too many errors. Below this level the data becomes too noisy or garbled
(unreadable).
• The receiver sensitivity is defined in several
complex ways but it basically refers to the
minimum allowable received power, causing the
minimum allowable CINR, in order to achieve a
specified BER.
It’s the minimum signal strength required at the receiver to get usable data without too many errors. Below this level the
data becomes too noisy or garbled (unreadable).

• The receiver sensitivity depends on noise floor,

interference cancellation techniques, how the
signal is modulated on the carrier, etc.
Modulation and Coding Scheme (MCS)

A user moves closer to the base station

Higher CINR (higher desired carrier power and lower interference)

Lower Error Vector Magnitude (EVM)

Adjustment of Modulation and Coding Scheme (MCS):

Higher Modulation level and/or higher code rate (e.g. 64-QAM in
place of 16-QAM; code rate ½ in place of 1/3 ).

Higher bit rate/ higher spectral efficiency

 Wrong Adjustment of MCS
A user moves further from the base station

Lower CINR (higher desired carrier power and lower interference)

Higher Error Vector Magnitude (EVM)

Wrong Adjustment of Modulation and Coding Scheme (MCS):

Higher Modulation level and/or higher code rate (e.g. 64-QAM in
place of 16-QAM; code rate ½ in place of 1/3 ).

Much higher packet retransmissions

Degrade in overall bit rate/spectral efficiency

Lower bit rate/ lower spectral efficiency
 Modulation Schemes Used in 3GPP
• GSM/GPRS uses GMSK which is a variant of FSK
• EDGE uses GMSK and 8-PSK
• UMTS uses supports QPSK.
• HSDPA supports QPSK and 16-QAM, HSUPA
supports BPSK, and HSPA+ additionally supports 64-
QAM in downlink and 16-QAM in uplink).
• Early releases of LTE supports QPSK, 16-QAM, and
64-QAM for data in both downlink and uplink.
– In Release 12, 256-QAM support is added for data in
downlink.
– In Release 14, 256-QAM support is added for data in uplink.
– In Release 15 (5G), 1024-QAM support is added for data in
downlink and uplink.
 Adaptive Modulation and Coding (AMC)
• If the link quality changes, Modulation and Coding Scheme
(MCS) needs to be adapted accordingly for more optimum bit
rate/ spectral efficiency.
• In AMC, the number of MCS levels has gradually increased with
the advancement of technology so that AMC can better tune to
radio link quality.
– GPRS uses GMSK with 4 MCS levels (in uplink and
downlink)
– EDGE uses GMSK and 8-PSK with 9 MCS levels (in uplink
and downlink)
– HSDPA uses QPSK and 16-QAM, with 16 MCS levels (in
downlink)
– LTE and 5G use QPSK, 16-QAM, and 64-QAM, 256-QAM
and 1024-QAM with 27 MCS levels (in uplink and downlink)
MCS Levels for Downlink in 5G
 Impact of SINR

If SINR gets better, the data rate improves because

• Higher Modulation and Coding Scheme (MCS)
can be used.
• Fewer packet retransmissions occur.
• The UE may get larger share of resource blocks
among users in the cell.
 Link Adaptation Techniques in 4G/5G
Link adaptation refers to a set of techniques where signal transmission
parameters are changed on the fly to better adjust to the changing radio
link conditions.
1. Adaptive Modulation and Coding (AMC): The AMC can be used
by varying Modulation and Coding Scheme (MCS) in time domain.
2. Channel Dependent Scheduling (CDS): The radio resources
allocation to different UEs are adapted (in both time and frequency
domains). The amount allocated also depends on radio link quality.
3. Transmission Power Control (TPC): The transmit power can be
adjusted in an attempt to maintain a desired level of SINR.
4. Multiple Antenna Schemes: The multiple antenna schemes are
adapted depending on the channel conditions.
MCS Selection for Uplink Data Transfer in 4G/5G

• The eNodeB can estimate uplink channel quality based

on transmission from UE and also, based on the error
rate in the uplink data transfer.
• The eNodeB selects one of the 27 MCS levels based on
its estimation.
• The eNodeB asks the UE to use the selected MCS level
for transmission.
 CQI for Downlink Data Transfer in 4G/5G
• The UE sends Channel Quality Indicator (CQI) as an
indication of the data rate which can be supported by the
downlink channel. The UE determines CQI to be reported
based on measurements of the downlink reference
signals. CQI is a 4 bit information with 15 CQI values. The
UE selects a CQI based on the estimation that the
Modulation and Coding Scheme (MCS) corresponding to
the CQI value is the highest possible MCS that will allow
the UE to decode transport blocks with error rate
probability not exceeding 10%.
• The eNodeB selects one of the 27 MCS levels based on
one of 15 CQI values.
• The eNodeB uses the selected MCS level and notifies the
UE of the selected MCS level, so the UE can retrieve data.
CQI for Downlink Data Transfer in 5G
Handling Packet Errors
 Types of Applications
• Real-Time (RT) applications
• NonReal-Time (NRT) applications

Sensitive Tolerant
Error NRT RT
Delay RT NRT

Applications
RT Voice, Live Streaming Video, Real Time Gaming
NRT Web browsing, Buffered Streaming Video, E-mail,
Chat, FTP, P2P File Sharing
Error and Delay for Different Apps
 Layer 2

Layer 2 has 3 sublayers:

1. Medium Access Layer (MAC)

2. Radio Link Control (RLC)
3. Packet Data Convergence Control (PDCP)
 Different Layers in 3G and 4G
Handling Packet Retransmissions
 Correction at MAC
MAC performs retransmissions for all types of data
traffic. MAC uses Hybrid ARQ (HARQ), which combines
FEC and ARQ, to minimize the number of
retransmissions. In this case, when the receiver fails to
decode the data block, it sends a NACK to the
transmitter, but it keeps bits from the failed attempt for
future use. If a NACK is received or a certain time elapses
without any feedback, the transmitter retransmits the
data block. Then the receiver HARQ soft-combines bits
from the previous attempt with the current
retransmission.
 Correction at RLC and TCP
RLC and TCP retransmissions use automatic repeat
request (ARQ) in which the receiver sends an
acknowledgment to indicate whether it has correctly
received the packet or not. Retransmissions in RLC and
TCP are typically performed for only nonreal-time
applications (e.g. buffered streaming video, web
browsing, e-mail, chat, FTP, P2P file sharing, etc.) since
they are error-sensitive and delay-tolerant. In this case,
RLC operates in acknowledged mode (AM). Conversely,
the real-time applications (e.g voice, live streaming video,
online gaming, etc.) are error-tolerant and delay-
sensitive, hence no retransmissions except from MAC. In
this case, RLC operates in unacknowledged mode (UM)
and RTP/UDP is employed instead of TCP.
Correction at Different Layers

MAC RLC Upper Layer

Protocols
RT HARQ Unacknowledged RTP/UDP/IP
Mode (UM)

NRT HARQ Acknowledged TCP/IP

Mode (AM)
 Correction at Different Layers
If the error correction fails in MAC, RLC AM attempts to
recover the error and similarly, if RLC AM fails, TCP
attempts to recover. The following points show that a
correction in a lower layer, among MAC, RLC and TCP, is
always preferable.
• A lower layer uses smaller data block size and so, the
retransmission requires less bandwidth.
• A lower layer uses smaller number of bits in
acknowledgement requiring less overhead.
• A lower layer is processed before the higher layers are.
If a lower layer can correct, then it is quicker.