wireless communication fundamentals part iipart iidtipper/2700/2700_slides3k.pdfphase-shift keying...
TRANSCRIPT
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