Wireless Communication FundamentalsWireless Communication FundamentalsPart IIPart IIPart IIPart II

David TipperAssociate ProfessorAssociate Professor

Graduate Telecommunications and Networking Program

U i it f Pitt b hUniversity of Pittsburgh

TelcomTelcom 2700 Slides 32700 Slides 3

Typical Wireless Communication System

SourceSource Encoder

ChannelEncoder

Modulator

Channel

Telcom 2700 2

DestinationSource Decoder

ChannelDecoder

Demod-ulator

Modulation and demodulation

digitald l ti

digitaldata analog

d l ti

analogbasebandsignal

radio transmitter

synchronization

digitaldataanalog

analogbasebandsignal

modulation modulation

radiocarrier

101101001 radio transmitter

Telcom 2700 3

decisiondemodulation

radiocarrier

101101001 radio receiver

Modulation

• Modulation – Converting digital or analog information to a

waveform suitable for transmission over a given mediummedium

– Involves varying some parameter of a carrier wave (sinusoidal waveform) at a given frequency as a function of the message signal

– General sinusoid

• A cos (2fCt + )Phase

Telcom 2700 4

– If the information is digital changing parameters is called “keying” (e.g. ASK, PSK, FSK)

Amplitude Frequency

Phase

Modulation

• Motivation– Smaller antennas (e.g., /4 typical antenna size)

• = wavelength = c/f , where c = speed of light, f= frequency.

• 3000Hz baseband signal => 15 mile antenna, 900 MHz => 8 cm3000Hz baseband signal 15 mile antenna, 900 MHz 8 cm

– Frequency Division Multiplexing – provides separation of signals

– medium characteristics

– Interference rejection

– Simplifying circuitry

• Modulation– shifts center frequency of baseband signal up to the radio carrier

• Basic schemes

Telcom 2700 5

• Basic schemes– Amplitude Modulation (AM) Amplitude Shift Keying (ASK)

– Frequency Modulation (FM) Frequency Shift Keying (FSK)

– Phase Modulation (PM) Phase Shift Keying (PSK)

Digital Transmission

• Wireless networks have moved almost entirely to digital modulation

• Why Digital Wireless?• Why Digital Wireless?–– Increase System Capacity (voice compression) more efficient Increase System Capacity (voice compression) more efficient

modulation modulation

–– Error control coding, equalizers, etc. => lower power neededError control coding, equalizers, etc. => lower power needed

–– Add additional services/features (SMS, caller ID, etc..)Add additional services/features (SMS, caller ID, etc..)

–– Reduce CostReduce Cost

I S it ( ti ibl )I S it ( ti ibl )

Telcom 2700 7

–– Improve Security (encryption possible)Improve Security (encryption possible)

–– Data service and voice treated same (3G systems)Data service and voice treated same (3G systems)

• Called digital transmission but actually Analog signal carrying digital data

Digital modulation Techniques

• Amplitude Shift Keying (ASK):– change amplitude with each symbol

– frequency constant

1 0 1

tq y

– low bandwidth requirements

– very susceptible to interference

• Frequency Shift Keying (FSK):– change frequency with each symbol

– needs larger bandwidth

t

1 0 1

t

1 0 1

Telcom 2700 8

• Phase Shift Keying (PSK):– Change phase with each symbol

– More complex

– robust against interference

1 0 1

t

Basic Digital Modulation Techniques

Telcom 2700 9

Amplitude-Shift Keying

• One binary digit represented by presence of carrier, at constant amplitude

Oth bi di it t d b b f• Other binary digit represented by absence of carrier

• where the carrier signal is Acos(2πfct)

• B = 2f f = input bit rate

ts tfA c2cos

0

1binary

0binary

Telcom 2700 10

• B = 2fb, fb = input bit rate

• Very Susceptible to noise

• Used to transmit digital data over optical fiber

Binary Frequency-Shift Keying (BFSK)

• Two binary digits represented by two different frequencies near the carrier frequency

– where f1 and f2 are offset from carrier frequency fc by equal but opposite amounts

ts tfA 12cos

tfA 22cos

1binary

0binary

Telcom 2700 11

fc by equal but opposite amounts

– B = 2([f2 – f1]/2 + fb)• Where fb = input bit rate

Phase-Shift Keying (PSK)

• Two-level PSK (BPSK)– Uses two phases to represent binary digits

ts tfA c2cos tfA c2cos

1binary 0binary

tfA 2cos 1binary

Telcom 2700 12

B = 2fb

tfA c2cos

tfA c2cos1binary 0binary

• Given any modulation scheme, it is possible to obtain its signal constellation.– Represent each possible signal as a

Signal Constellation

p p gvector in a Euclidean space.

• In symbol detection – decode incoming signal as closest symbol in the signal constellation space

• If we know the signal constellation, we can estimate the performance in terms of the probability of symbol error given the noise

Telcom 2700 13

symbol error given the noise parameters.

• Probability of error depends on the minimum distance between the constellation points.

Advanced Modulation Schemes

• Variations on ASK, FSK and PSK possible

• Attempt to improve performance– Increase data for a fixed bandwidth

– Remove requirement for clock recovery

– Improve BER performance

• Main schemes for wireless systems are

Telcom 2700 14

• Main schemes for wireless systems are based on FSK and PSK because they are more robust to noise

M-ary Signaling/Modulation

• What is M-ary signaling?– The transmitter considers ‘k’ bits at a times. It

d f M i l h M 2kproduces one of M signals where M = 2k.

Example: Quarternary PSK (k = 2)

Tttf c 0 , 2cos 00 T2E

Input: Signal :

Telcom 2700 18

Tttf

Tttf

Tttf

c

c

c

0 ,2cos 10

0 , 2cos 11

0 , 2cos 01

23

T2E

T2E

2T2E

QPSK Constellations

( ) ( )

1(t)

2(t) 2(t)

1(t)

Telcom 2700 19

Rotated by /4

QPSK

Telcom 2700 20

M-ary Error Performance

A received symbol is decoded into the closest the symbol in the signal constellation

BPSK

signal constellation

As the number of symbols M in the signal space increases the decoding region for each symbol decreases BER goes up

•For example MPSK as M QPSK

Telcom 2700 21

For example MPSK, as M increases

–the bandwidth remains constant,

–increase in symbol error rate

Q S

Selection of Encoding/Modulation Schemes

• Performance in an Noisy channel

– How does the bit error rate vary with the energy per bit

available in the systemavailable in the system

• Performance in fading multipath channels

– Same as above, but add multipath and fading

• Bandwidth requirement for a given data rate

– Also termed spectrum efficiency or bandwidth efficiency

H bit / i H f b d idth

Telcom 2700 23

– How many bits/sec can you squeeze in one Hz of bandwidth

• Cost

– The modulation scheme needs to be cost efficient

Performance in Noisy channels

• AWGN = Additive White Gaussian Noise– This has a flat noise

spectrum with average10

0Binary modulation schemes

spectrum with average noise power of N0

• The probability of bit error (bit error rate) is measured as a function of ratio of the “energy per bit” – Eb to the average noise value– BER or Pe variation with

E /N10

-4

10-3

10-2

10-1

Pe

BPSKDPSKBFSK

Telcom 2700 24

Eb/N0

– Eb/N0 is a measure of the “power requirements”

• Tradeoffs!• Some form of PSK used in

most wireless systems

2 4 6 8 10 12 1410

-6

10-5

Eb/N

0

Effect of Mobility?

• A moving receiver can experience a positive or negative Doppler shift in received signal, depending on direction of movement– Results in widening frequency spectrum

• Fading is a combination of fast fading (Short term fading and Intersymbol interference) and long term fading (path loss- shadowing)

Telcom 2700 25

[Garg and Wilkes Fig 4.1]

Performance in Fast Fading Channels

• The BER is now a function of the “average” Eb/N0

100

Binary modulation schemes under fading

BPSK-No fadingg b 0

• The fall in BER is linear

• Large power consumption on average to achieve a good BER– 30 dB is three orders of

magnitude larger

• Must use Diversity 10-4

10-3

10-2

10-1

Pe

BPSKDPSKBFSK

Telcom 2700 26

Must use Diversity techniques to overcome effect of Short Term (Fast) Fading and Multipath Delay

5 10 15 20 25 30 35 40 4510

-6

10-5

10

Eb/N

0

What is Diversity?

• Idea: Send the same information over several “uncorrelated” forms– Not all repetitions will be lost in a fade

T f di it• Types of diversity– Time diversity – repeat information in time spaced so as to

not simultaneously have fading• Error control coding!

– Frequency diversity – repeat information in frequency channels that are spaced apart

• Frequency hopping spread spectrum

Space diversity use multiple antennas spaced sufficiently

Telcom 2700 27

– Space diversity – use multiple antennas spaced sufficiently

apart so that the signals arriving at these antennas are not

correlated• Usually deployed in all base stations but harder at the mobile

Performance Degradation and Diversity

Issue Performance Affected DiversityTechniqueq

ShadowFading

Received Signal Strength Fade Margin – Increasetransmit power

or decrease cell size

Fast Fading

Bit error ratePacket error rate

Antenna DiversityError control coding

InterleavingF eq enc hopping

Telcom 2700 28

Frequency hopping

MultipathDelaySpread

Inter-symbol InterferenceAdaptive EqualizationDS-Spread Spectrum

OFDMDirectional Antennas

Error Control

• BER in wireless networks – Several orders of magnitude worse than wireline

networks (eg 10-2 vs 10-10 in optic fibers)networks (eg, 10 vs 10 in optic fibers)

– Channel errors are random and bursty, usually coinciding with deep fast fades

– Much higher BER within bursts

• Protection against bit errors – Necessary for data

Telcom 2700 29

y

– Speech can tolerate much higher bit errors (< 10-2

depending on encoding/compression algorithm)

• Error Control Coding used to overcome BER

Error control coding

• Coding is a form of diversity

Simple Block codey

– Transmit redundant bits from which you can detect/recover from errors

– The redundant bits have a pattern that enables this recovery

BlockEncoder

k bit data block n bit codeword

Telcom 2700 30

• Approaches to error control– Error Detection + ARQ– Error Correction (FEC) n – k parity check bits k data bits

Error control

• Error control coding: systematically add redundant bits for error detection or correction– Error detection codes:

• Detect whether received word is a valid “codeword” but not enough redundancy to correct bits

• Retransmit data after error (automatic repeat request –ARQ)

– Error correction codes (forward error correction: FEC)• Detect invalid codewords and correct into valid codeword

Telcom 2700 31

• Correction requires more bits than error detection• FEC is good for one-way channels, recordings (CD-ROMs),

real-time communications, deep space,...

– Generally more bits are required to protect against larger number of bit errors

Single Parity

• Example: single parity bit even parity code– Valid codewords should always have even

number of 1’s

– Add a parity bit=1 if number of 1’s in data is odd add parity bit=0 if number of 1’s in data is even

– If any bit is errored, the received codeword will have odd number of 1’s

Telcom 2700 32

a e odd u be o s

– Single parity can detect any single bit error (but not correct)

• Actually, any odd number of bit errors can be detected

Single Parity (cont)

Examplereceived codewords

3 bits parity bit

0000 00010011 00100101 01000110 01111001 10001010 10111100 11011111 1110

valid invalidExample

--> 23 valid codewords

transmission

Telcom 2700 33

Single bit error will change valid word into an invalid word (detectable);double bit error will change valid wordinto another valid word (undetectable)

Block Codes

• (n,k) block codesk = number of data bits in block (data word (

length)

n-k = number of parity check bits added

n = length of codeword or code block

(n-k)/n = overhead or redundancy (lower is more efficient)

Telcom 2700 34

(lower is more efficient)

k/n = coding rate (higher is more efficient)

Block Codes (cont)

data • • •• • •

data

k bits

parity check

data

n bits

--> 2k valid codewords

n-bit codeword

2n possible codewords,only 2k are valid

transmission

Telcom 2700 35

valid?

no bit errors

error

Block Code Principles

• Hamming distance :– for 2 n-bit binary sequences, the number of different bitsfor 2 n bit binary sequences, the number of different bits

– e.g., v1=011011; v2=110001; d(v1, v2)=3

• The minimum distance (dmin) of an (n,k) block code is the smallest Hamming distance between any pair of codewords in a code. – Number of error bits can be detected: dmin-1

Telcom 2700 36

min

– Number of error bits can be corrected t:

2

1mindt

(7,4) Hamming code

Message word Code word Weight

0000 000 0000 0

0001 101 0001 3

0010 111 0010 4

0011 010 0011 3

0100 011 0100 3

0101 110 0101 4

0110 100 0110 3

0111 001 0111 4

1000 110 1000 3

1001 011 1001 4

27 possible7-bit words (128 possible) of which weuse only 16

All codewords are distance 3 apart > Can

Telcom 2700 37

1010 001 1010 3

1011 100 1011 4

1100 101 1100 4

1101 000 1101 3

1110 010 1110 4

1111 111 1111 7

apart => Can detect 2 errors, correct 1 error

Forward Error Correction Process

FEC Operation•Transmitter

Forward error correction (FEC)–Forward error correction (FEC) encoder maps each k-bit block into an n-bit block codeword–Codeword is transmitted;

•Receiver–Incoming signal is demodulated–Block passed through an FEC decoder

Telcom 2700 38

decoder–Decoder detects and correct errors

• Receiver can correct errors by mapping invalid codeword to nearest valid codeword

FEC (cont)

Decoding

valid codewordC1

valid codewordC2

distance d 2e+1

distance e

Decodingsphere

Telcom 2700 39

distance e

Convolutional Codes

• Block codes treat data as separate blocks (memoryless encoding);

• Convolutional codes map a continuous data string into aConvolutional codes map a continuous data string into a continuous encoded string (memory)

• Error checking and correcting carried out continuously– (n, k, K) code

• Input processes k bits at a time • Output produces n bits for every k input bits• K = constraint factor

Telcom 2700 40

• k and n generally very small

– n-bit output of (n, k, K) code depends on:• Current block of k input bits• Previous K-1 blocks of k input bits

Convolutional Codes (cont)

Successive k-tuples are mapped into n-tuples • n-tuples should be designed to have distance

properties for error detection/correction

• Example: K=3 stages, k=1, n=2 bits output

bitinput data bit bit

Telcom 2700 41

+ +

first bit second bit

Convolutional Encoder

Can represent coder by state transition diagram – with 2(K-1) states

Telcom 2700 42

What does coding get you?

• Consider a wireless link – probability of a bit error = q p y q

– probability of correct reception = p

– In a block of k bits with no error correction

– P(word correctly received) = pk

– P(word error) = 1 – pk

i h i f bi i bl k f bi

Telcom 2700 43

– With error correction of t bits in block of n bits

) (1) (

)() (0

correctwordPerrorwordP

qpi

ncorrectwordP iin

t

i

What does coding get you?

• Example consider (7,4) Hamming Code when BER = q = .01, p = .99– In a block of 4 bits with no error correction– P(word correctly received) = pk = .9606– P(word error) = 1 – pk = 0.04– With error correction of 1 bits in block of 7 bits

998.0)(6

7)() ( 177

0

qppqp

i

ncorrectwordP iin

t

Telcom 2700 44

– Get an order of magnitude improvement in word error rate

002.0) (1) (

60

correctwordPerrorwordP

ii

Interleaving

• Problem: – Errors in wireless channels occur in bursts due to fast fades

– Error correction codes designed to combat random errors inError correction codes designed to combat random errors in the code words

• Interleaving Idea:– If the errors can be spread over many codewords they can be

corrected- achieved my shuffling codewords

– makes the channel memoryless and enables coding schemes to perform in fading channels.

f

Telcom 2700 45

– the penalty is the delay in receiving information- bits have to be buffered for interleaving

– Interleaving is performed after coding – at receiver de-interleave before decoding

Block interleaving

• After codewords are created, the bits in the codewords are interleaved and transmittedThis ensures that a burst of

codeword

• This ensures that a burst of errors will be dispersed over several codewords and not within the same codeword

• Needs buffering at the receiver to create the original data

Bits in error

dispersed over

several codewords

Telcom 2700 46

• The interleaving depth depends on the nature of the channel, the application under consideration, etc.

Direction of transmission

Interleaving Example

• Usually transmit data in order it arrives

18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 :positionbit

• Suppose bits 6 to 11 are in error because of a fade The codewords a and b are lost.

• Suppose we interleaving at depth 7 by buffering up 7 words then output them in order to bit positions of the words

321065432106543210 c c c c b b b b b b b a a a a a a a :bit

181716151413121110987654321itibit

Telcom 2700 47

• Now can correct a fade that results in bits 6-11 being lost

222211111110000000 d c b a g f e d c b a g f e d c b a :bit

18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 :positionbit

Frequency Hopping

• Traditionally: transmitter/receiver pair communicate on fixed frequency channel.Frequency Hopping Idea:• Frequency Hopping Idea:– Since noise, fading and interference

change somewhat with frequency band used – move from band to band

– Time spent on a single frequency is termed the dwell time

• Originally for military communicationsSpend a short amount of time on different

Telcom 2700 48

– Spend a short amount of time on different frequency bands to prevent interception or jamming - developed during WWII by actress Hedy Lammar and classical composer George Antheil – patent given to government

Frequency Hopping concept

f

f3

f4

f5

f6

f7

f8

Telcom 2700 49

time

Timeslot for transmission

f1

f2

f3

Unacceptable errors

Frequency Hopping

• Two types:– Slow Hopping

• Dwell time long enough to transmit several bits in a row (timeslot)

– Fast Hopping• Dwell time on the order of a bit or fraction of a bit (primarily for

military systems)

• Transmitter and receiver must know hopping pattern/ algorithm before communications.– Cyclic pattern – best for low number of frequencies and

combating Fast Fading : E l ith f f i f4 f2 f1 f3 f4 f2 f1 f3

Telcom 2700 50

• Example with four frequencies: f4, f2, f1, f3, f4, f2, f1, f3, ….

– Random pattern – best for large number of frequencies, combating co-channel interference, and interference averaging

• Example with six frequencies: f1, f3, f2, f1, f6, f5, f4, f2, f6, … • Use random number generator with same seed and both ends

Frequency Hopping

• Slow frequency hopping used in cellular (GSM)( )

• Fast in WLANs

• Provides interference averaging and frequency diversity

• By hopping mobile less like to suffer consecutive

Telcom 2700 51

like to suffer consecutive deep fades

Multipath propagation

• Signal can take many different paths between sender and receiver due to reflection, scattering, diffraction

signal at sendersignal at receiver

• Time dispersion: signal is dispersed over time• interference with “neighbor” symbols Inter Symbol

Telcom 2700 52

interference with neighbor symbols, Inter Symbol Interference (ISI)

• The signal reaches a receiver directly and phase shifted• distorted signal depending on the phases of the different

parts

Equalization

• Equalizer– filter that performs the inverse of the channel to compensate for

the distortion created by multipath delay (combats ISI)

• In wireless networks equalizers must be adaptive• In wireless networks equalizers must be adaptive– channel is usually unknown and time varying– equalizers track the time variation and adapt

• Two step approach to equalization1. Training:

• a known fixed-length sequence is transmitted for the receiver’s equalizer to ‘train’ on – that is to set parameters in the equalizer

2 Tracking:

Telcom 2700 53

2. Tracking:• the equalizer tracks the channel changes with the help of the

training sequence, and uses a channel estimate to compensate for distortions in the unknown sequence.

Adaptive Equalization

• Training step, the channel response, h(t) is estimated• Tracking step, the input signal, s(t), is estimated

• Equalizers are used in NA-TDMA, GSM and HIPERLAN• There are several different types of equalizers (3 popular ones)

Telcom 2700 54

There are several different types of equalizers (3 popular ones)• LTE: Linear transversal filter• DFE: Decision feedback equalizer• MLSE: Maximum likelihood sequence estimators

• Disadvantages of equalizers are complexity, bandwidth consumption and power consumption

Adaptive Equalizers

Typical equalizer implemented as a tap delay line filter with variable tap gains

Telcom 2700 55

Spread Spectrum Radio Aspects

• Military Spread spectrum techniques adapted for cellular systems 1. Frequency Hopping: vary frequency transmit on 2. Direct Sequence

• Narrowband signal is spread over very large bandwidth signal using a spreading signal

• Spreading signal is a special code sequence with rate much greater than data rate of message

• The receiver uses correlation to recover the original data

Telcom 2700 56

• Multipath fading is reduced by direct sequence signal spreading and better noise immunity

• DS also allows lower power operation

Direct Sequence Spread Spectrum

• Each bit in original signal is represented by multiple bits in the transmitted signal

• Spreading code spreads signal across a wider frequency band – Spread is in direct proportion to number of chip bits W used– Processing gain G = W/R; W = chips per sec, R = information bit

rate per sec– Processing gain is a measure of the improvement in SNR gained

by using the additional bandwidth from spreading (18-23 dB in ll l )

Telcom 2700 57

cellular systems)

• One Spreading technique combines digital information stream with the spreading code bit stream using exclusive-OR

DSSS Modulation

• The original data stream is “chipped” up into a pattern of pulses of smaller duration

Data Bit Data In

• Good correlationproperties

• Good cross-correlation properties with other patterns

• Each pattern is called a spread spectrum code

• Instead of transmit 1

“Spread” Bits

1 2 3 4 5 6 7 8 9 10 11

SpreadingCode In

Telcom 2700 58

• Instead of transmit 1 data bit transmit 11 chip bit pattern - adding redundancy

chip

Periodic Spreading Code

Code In

DSSS (Direct Sequence Spread Spectrum) II

Xuser data

modulator

spreadspectrumsignal

transmitsignal

chippingsequence

radiocarrier

transmitter

demodulator

receivedsignal

X

lowpassfilteredsignal

integrator

products

decisiondata

sampledsums

correlator

Telcom 2700 59

demodulator

radiocarrier

X

chippingsequence

receiver

integrator decision

DSSS Spectrum

1

•Example: IEEE 802.11 Wi-Fi Wireless LAN standardUses DSSS with 11 bit chipping

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1original spectrumspread spectrum

am

plit

ude

Uses DSSS with 11 bit chipping code

•To transmit a “0”, you send[1 1 1 -1 -1 -1 1 -1 -1 1 -1]

•To transmit a “1” you send [-1 -1 -1 1 1 1 -1 1 1 -1 1]

•Processing gain–The duration of a chip is usually represented by Tc

Telcom 2700 60

-10 -8 -6 -4 -2 0 2 4 6 8 100

0.1

0.2

spectrum

normalized frequency

y c

–The duration of the bit is T

–The ratio T/Tc = R is called the “processing gain” of the DSSS system–For 802.11 R = 11

Example in a two-path channel

• Random data sequence of ten data bits– Spreading by 11 chips using

1.5

2

0 0 0 0 1 1 0 0 1 0

– Spreading by 11 chips using 802.11b chipping code

• Two path channel with inter-path delay of 17 chips > bit duration

• Multipath amplitudes– Main path: 1

-2

-1.5

-1

-0.5

0

0.5

1

Telcom 2700 61

– Second path: 1.1

• Reality: – Many multipath components– Different path amplitudes– Noise

10 20 30 40 50 60 70 80 90 100 1102

Output without spreading

Errorsintroduced by the channel

-5

0

5

10

15

20

25

-5

0

5

10

15

the channel

0 0 0 0 1 1 0 0 1 00 0 0 0 1 1 0 0 1 0

Telcom 2700 62

0 20 40 60 80 100 120-15

-10

0 20 40 60 80 100 120-15

-10

Without Multipath With Multipath

Output with spreading

Output of DSS demodulator

Errors introduced by the channel are removed

-5

0

5

10

15

p

-5

0

5

10

15

0 0 0 0 0 0011 1 0 0 0 0 0 0011 1

Telcom 2700 63

0 20 40 60 80 100 120-15

-10

Without Multipath With Multipath

0 20 40 60 80 100 120-15

-10

Performance Degradation and DiversityIssue Performance Affected Diversity

Technique

Shadow Received Signal Strength Fade Margin – IncreaseFading transmit power

or decrease cell size

Fast Fading(Time Variation)

Bit error ratePacket error rate

Antenna DiversityError control coding

InterleavingFrequency hopping

Telcom 2700 64

MultipathDelaySpread

(Time Dispersion)

Inter-symbol InterferenceAdaptive EqualizationDS-Spread Spectrum

OFDMDirectional Antennas

Typical Wireless Communication System

SourceSource Encoder

ChannelEncoder

Modulator

Channel

Telcom 2700 65

DestinationSource Decoder

ChannelDecoder

Demod-ulator

Source Coding

• Source Coding seeks to efficiently encode the information for transmission

• Mobile environment much different from wired – issues– Efficient use of spectrum

• Compress to lower bit rate per user => more users

– Quality • Want near tollgrade in a very difficult transmission environment

(high BER, bursty errors)

– Hardware complexity• Size speed and power consumption

Telcom 2700 66

Size, speed and power consumption

• Consider Digital Speech as example– Convert speech to digital form and transmit digitally

Digital SpeechDigital Speech

• Speech Coder: device that converts speech to digital• Types of speech coders

– Waveform coders– Waveform coders• Convert any analog signal to digital form

– Vocoders (Parametric coders)• Try to exploit special properties of speech signal to reduce bit rate• Build model of speech – transmit parameters of model

– Hybrid Coders• Combine features of waveform and vocoders

Telcom 2700 67

Waveform Coders (e.g.,PCM)• Waveform Coders

• Convert any analog signal to digital -basically A/D converter• Analog signal sampled > twice highest frequency- then quantized into ` n ‘ bit samples • Uniform quantization • Example Pulse Code Modulation

• band limit speech < 4000 Hz• pass speech through law compander • sample 8000 Hz, 8 bit samples • 64 Kbps DS0 rate

Telcom 2700 68

64 Kbps DS0 rate

• Characteristics• Quality – High• Complexity – Low• Bit rate – High• Delay - Low• Robustness - High

• Bandwidth – Most of energy between 20 Hz to

about 7KHz ,

Characteristics of SpeechCharacteristics of Speech

– Human ear sensitive to energy between 50 Hz and 4KHz

• Time Signal– High correlation– Short term stationary

• Classified into four categories– Voiced : created by air passed

Telcom 2700 69

y pthrough vocal cords (e.g., ah, v)

– Unvoiced : created by air through mouth and lips (e.g., s, f )

– Mixed or transitional– Silence

Typical Voiced

h

Characteristics of Speech

speech

Typical U i d

Telcom 2700 70

Unvoiced speech

• Vocoders (Parametric Coders)• Models the vocalization of speech

•Speech sampled and broken into frames (~25 msec)

Vocoders

• Instead of transmitting digitized speech build model of speech –transmit parameters of model and synthesize approximation of speech

• Linear Predictive Coders (LPC)

• Models Vocal tract as a filter

•Filter excitation

• periodic pulse (voiced speech)

• noise (unvoiced)

Telcom 2700 71

• Transmitted parameters: gain, voiced/unvoiced decision, pitch (if voiced), LPC parameters

• Example Tenth Order Linear Predictive Coder

• Samples Voice at 8000 Hz – buffer 240 samples => 30 msec

• Filter Model

Vocoders

• (M=10 is order, G is gain, z-1 unit delay, bk are filter coefficients)

• G = 5 bits, bk = 8 bits each, voiced/unvoiced decision = 1 bit, pitch 6 bit 92 bit /30 3067 b

kk

M

k

zb

GzH

1

1)(

Telcom 2700 72

= 6 bits => 92 bits/30 msec = 3067 bps

Vocoders

• LPC coders can achieve low bit rates 1.2 – 4.8 Kbps

• Characteristics of LPCQ lit L• Quality – Low

• Complexity – Moderate• Bit Rate – Low• Delay – Moderate• Robustness – Low

• Quality of pure LPC vocoder to low for cellular telephony -try to improve quality by using hybrid coders

Telcom 2700 73

y p q y y g y

• Hybrid Coders• Combine Vocoder and Waveform Coder concept

• Residual LPC (RELP)• Codebook excited LPC (CELP)

• Residual Excited LPC• improve quality of LPC by transmitting error (residue) along with LPC parameters

RELP Vocoder

Buffer/Window

ST-LPAnalysis

EN

CO

DE

R

residue

LP parameters

Encodedoutput

v/u decisiongain, pitch

s(n)

Telcom 2700 74

Block diagram of a RELP encoder

LPCSynthesis

g , p

GSM Speech Coding

Low-passfilter

Analogspeech

A/DRPE-LTPspeechencoder

Channelencoder

8000 samples/s,13 bits/sample

104 kbps 13 kbps

Telcom 2700 75

13 bits/sample

GSM Speech Coding (cont)

Regular pulse excited - long term prediction (RPE-LRP)speech encoder (RELP speech coder

RPE-LTPspeechencoder

160 samples/20 ms from A/D

(= 2080 bits)

36 LPC bits/20 ms9 LTP bits/5 ms47 RPE bits/5 ms

260 bits/20 msto channelencoder

Telcom 2700 76

LPC: linear prediction coding filterLTP: long term prediction – pitch + inputRPE: Residual Prediction Error:

Hybrid Vocoders

• Codebook Excited LPC• Problem with simple LPC is U/V decision and pitch estimationdecision and pitch estimation doesn’t model transitional speech well, and not always accurate

•Codebook approach – pass speech through an analyzer to find closest match to a set of possible excitations (codebook)

Telcom 2700 77

excitations (codebook)

• Transmit codebook pointer + LPC parameters • 3G, ITU G.729 standard

CELP Speech Coders

• General CELP architecture

Telcom 2700 78

Evaluating Speech Coders

• Qualitative Comparison– based on subjective procedures

in ITU-T Rec. P. 830

• Major Procedures

Mean Opinion Score (MOS)-------------------------------------

Excellent 5• Major Procedures• Absolute Category Rating

– Subjects listen to samples and rank them on an absolute scale - result is a mean opinion score (MOS)

• Comparison Category Rating– Subjects listen to coded

Excellent 5Good 4Fair 3Poor 2Bad 1

Comparison MOS (CMOS)-------------------------------------

Telcom 2700 79

jsamples and original un-coded sample (PCM or analog), the two are compared on a relative scale – result is a comparison mean opinion score (CMOS)

Much Better 3Better 2Slightly Better 1

About the Same 0Slightly Worse -1Worse -2Much Worse -3

Evaluating Speech Coders

MOS for clear channel environment – no errorsResult vary a little with language and speaker gender

Standard Speech coder Bit rate MOSPCM Waveform 64 Kbps 4.3

CT2 Waveform 32 Kbps 4.1DECT Waveform 32 Kbps 4.1NA-TDMA Hybrid CELPC 8Kbps 3.0

GSM Hybrid RELPC 13 kbps 3.54

Telcom 2700 80

QCELP Hybrid CELPC 14.4 Kbps 3.4 – 4.0

QCELP Hybrid CELPC 9.6 Kbps 3.4

LPC Vocoder 2.4 Kbps 2.5ITU G.729 Hybrid CELP 8.Kbps 3.9

Multiple Access and Mode

• Mode – Simplex – one way communication (e.g., broadcast AM)

D l t i ti– Duplex – two way communication• TDD – time division duplex – users take turns on the channel

• FDD – frequency division duplex – users get two channels – one for each direction of communication

– For example one channel for uplink (mobile to base station) another channel for downlink (base station to mobile)

• Multiple Access determines how users in a cell share the frequency spectrum assigned to the cell: FDMA

Telcom 2700 81

the frequency spectrum assigned to the cell: FDMA, TDMA, CDMA

• Wireless systems often use a combination of schemes; GSM – FDD/FDMA/TDMA

Multiple Access Techniques

• FDMA (frequency division multiple access) – separate spectrum into non-overlapping frequency bands

assign a certain frequency to a transmission channel– assign a certain frequency to a transmission channel between a sender and a receiver

– different users share use of the medium by transmitting on non-overlapping frequency bands at the same time

• TDMA (time division multiple access): – assign a fixed frequency to a transmission channel between

a sender and a receiver for a certain amount of time (users

Telcom 2700 82

a sender and a receiver for a certain amount of time (users share a frequency channel in time slices)

• CDMA (code division multiple access): – assign a user a unique code for transmission between

sender and receiver, users transmit on the same frequency at the same time

Frequency division multiple access

ncy

freq

ue

Telcom 2700 time

Time Division Multiple Access

frameslot

freq

uenc

y

Telcom 2700 84

time

Code Division Multiple Accesscode

Telcom 2700 85

time

frequency

FDMA

• FDMA is simplest and oldest method• Bandwidth F is divided into T non-overlapping frequency

channelschannels– Guard bands minimize interference between channels– Each station is assigned a different frequency

• Can be inefficient if more than T stations want to transmit or traffic is bursty (resulting in unused bandwidth and delays)

• Receiver requires high quality filters for adjacent

Telcom 2700 86

channel rejection• Used in First Generation Cellular (NMT)

f1 f2

FDD/FDMA - general scheme, example AMPS (B block)

f

799893.97MHz 799

355

799

20 MHz

30 kHz880.65 MHz

849.97 MHz

Telcom 2700 87

t

355835.65 MHz

f(c) = 825,000 + 30 x (channel number) KHz <- uplinkf(c) = f uplink + 45,000 KHz <- downlinkIn general all systems use some form of FDMA

TDMA

• Users share same frequency band in non-overlapping time intervals, eg, by round robin

• Receiver filters are just windows instead of bandpass filters (as in FDMA)

• Guard time can be as small as the synchronization of the network permits– All users must be synchronized with base station

to within a fraction of guard time

Telcom 2700 88

to within a fraction of guard time– Guard time of 30-50 microsec common in TDMA

• Used in GSM, NA-TDMA, (PDC) Pacific Digital Cellular

TDD/TDMA - general scheme, example

1 2 3 11 12 1 2 3 11 12

tdownlink uplink

417 µs

Telcom 2700 89

CT2 cordless phone standard

Air Interface

• 25 MHz of bandwidth is divided into 124 frequency bands of 200 kHz each and two 100 kHz pieces on either side

• Carrier frequencies are given by:– Fu (n) = 890.2 + 0.2(n-1) MHz n=1,2,3,…,124– Fd (n) = 935.2 + 0.2(n-1) MHz n=1,2,3,…,124

• Example: On the uplink Channel 1 = 890 1 890 3 MHz

Telcom 2700 90

– On the uplink, Channel 1 = 890.1-890.3 MHz– On the downlink, Channel 1 = 935.1-935.3 MHz

• Usually, Channels 1 and 124 will not be used if possible

CDMA

• Code Division Multiple Access– Narrowband message signal is multiplied by very large bandwidth

di i l i di dspreading signal using direct sequence spread spectrum

– All users can use same carrier frequency and may transmit simultaneously

– Each user has own unique access spreading codeword which is approximately orthogonal to other users codewords

– Receiver performs time correlation operation to detect only specific codeword, other users codewords appear as noise due to

Telcom 2700 91

decorrelation

– Cocktail party example

CDMA and direct sequence spread spectrum

• The original d t t i Data BitVdata stream is “chipped” up into a pattern of pulses of smaller duration

• Each pattern is called a spread spectrum code

• The primary 0.3

0.4

0.5

0.6

0.7

0.8

0.9

1original spectrumspread spectrum

Data Bit

“Spread” Bit

1 2 3 4 5 6 7 8 9 10 11

V

V

t

am

plit

ud

e

Telcom 2700 92

p yspread spectrum code used in CDMA is the “Walsh Code”

-10 -8 -6 -4 -2 0 2 4 6 8 100

0.1

0.2

0.3

chip

t

normalized frequency

Simple example illustrating CDMA

Traditional • To send a 0, send +1 V for T

seconds

Simple CDMA• To send a 0, Bob sends +1 V for Tseconds

• To send a 1, send -1 V for T seconds

• Use separate time slots or frequency bands to separate signals

To send a 0, Bob sends 1 V for T seconds; Alice sends +1 V for T/2 seconds and -1 V for T/2 seconds

• To send a 1, Bob sends -1 V for T seconds; Alice sends -1 V for T/2 seconds and +1 V for T/2

Telcom 2700 93

1 01

[1, 1] [-1, -1][1, -1] [-1, 1]

Code

chip

T T T T

11

time

V

Data

0

Simple CDMA Transmitter

SpreadUser 1 data in

SpreadUser 2 data in

0 0 1 1 = [1 1 1 1 -1 -1 -1 -1]

Telcom 2700 94

0 0 1 1

1 0 1 0

= [1, 1, 1, 1, -1, -1, -1, -1]

= [-1, 1, 1, -1, -1, 1, 1, -1]

Transmitted signal

t

V

V

t

t

V

T 2T 3T 4T

T 2T 3T 4T

T 2T 3T 4T

1

1

2

Simple CDMA Receiver

Despread User 1 data out

Despread User 2 data out

Received

V

correlate with [1, 1]

correlate with [1, -1]

2

Telcom 2700 95

Received signal t

t

V

Alice’sCode =

= -2 x T/2 = -T

-T has a negative sign Alice sent a 1

as the first bit

T 2T 3T 4T

T 2T 3T 4T

1

Simple CDMA continued

• Proceeding in this fashion for each “bit”, the information transmitted by Alice can be recovered

• To recover the information transmitted by Bob, theTo recover the information transmitted by Bob, the received signal is correlated bit-by-bit with Bob’s code [1,1]

• Such codes are “orthogonal”– Multiply the codes element-wise

• [1,1] x [1,-1] = [1,-1]– Add the elements of the resulting product

• 1 + (-1) = 0 => the codes are orthogonal

I CDMA (IS 95) h th l d i ll d

Telcom 2700 96

• In CDMA (IS-95) each orthogonal code is called a Walsh Code and has 64 chips in it

• Note that the transmissions MUST be synchronized to recover data– Otherwise there will be interference and errors

CDMA Properties: Near-Far Problem

• A CDMA receiver cannot successfully despread the desired signal in a high multiple-access-interference environmentenvironment

• Unless a transmitter close to the receiver transmits at power lower than a transmitter farther away, the far transmitter cannot be heard

Telcom 2700 97

Base station

• Power control must be used to mitigate the near-far problem

• Mobile transmit so that power levels are equal at base station

Summary

• Fundamental Concepts for Wireless Communications– Modulation Techniques

– Digital Modulation

Di i T h i• Diversity Techniques– Error Control Coding

– Interleaving

– Adaptive Equalizers

– Frequency Hopping Spread Spectrum

– Direct Sequence Spread Spectrum

M d d M lti l A

Telcom 2700

• Mode and Multiple Access– TDD/FDD

– FDMA

– TDMA

– CDMA