DigComm_Fall09-Chapter4

Post on 23-Oct-2014

113 views 1 download

Tags:

Transcript of DigComm_Fall09-Chapter4

Chapter Four:

Bandpass Modulation

What is bandpass modulation? Baseband modulation

The form of shaped pulses

Bandpass modulation The shaped pulses modulate a sinusoid, called a carrier

wave or a carrier Types

Phase shift keying Frequency shift keying Amplitude shift keying Continuous phase modulation Hybrid

Why using a carrier? The carrier converts to an electromagnetic field for

propagation through antennas The size of the antenna depends on the wavelength λ

Ex: Cellular telephone, antenna size λ/4

cm

MHzf

milesm

m

Hzf

smcf

c

84

10900/100.3 bemight size antenna The

,900carrier aWith

15105.24

bemight size antenna The

103000

100.3

,3000 signal Baseband

/100.3, Wavelength

68

4

58

8

Sinusoidal Waveforms The general form of the carrier wave

)(cos)()( ttAts

Time-varying amplitude Time-varying angle

)(cos)()( 0 tttAts

Radian frequency 2πf Phase

Coherent vs. non-coherent With/without the knowledge of the carrier’s phase to detect the signals Complexity vs. performance

Phase Shift Keying (PSK)

M

it

MiTtttT

Ets

i

ii

2)(

,...2,1,0,)(cos2

)( 0

duration symbol :energy symbol : TE

Waveform Vector

s2s1

M=2

T T T

Frequency Shift Keying (FSK)

MiTttT

Ets ii ,...2,1,0,cos

2)(

duration symbol :energy symbol : TE

Waveform Vector

s2

s1

M=3

s3T T T

Frequency Shift Keying (FSK) Orthogonal signals in FSK

Not all FSK signaling is orthogonal Example: f1=10000Hz, f2=11000Hz, are they

orthogonal?? To meeting the criterion the spacing between

the tones [on page 202, example 4.3]

A frequency separation is a multiple of 1/T Hz The minimum requirement spacing for noncoherent

detection is 1/T The minimum requirement spacing for coherent

detection is 1/(2T)

Amplitude Shift Keying (ASK)

MiTttT

tEts i

i ,...2,1,0,cos)(2

)( 0

Waveform Vector

Binary ASK (On-Off Keying)

s2 s1

M=2

T T T

Hybrid -- Amplitude Phase Keying

MiTtttT

tEts i

ii ,...2,1,0,)(cos

)(2)( 0

Waveform Vector

T

T

T M=8

Where are we?!

Format Pulse modulation Bandpass modulation

Format Demodulation, sampling and detection

Correlation Receiver

MiTttntstr i ...2,1,0)()()(

Receiver:

Step 1: reduce r(t) to a single random variable z(T) or a set of RVs zi(T)

Step 2: a symbol decision is made on the basis of comparing z(T) to a threshold or choosing the max zi(T)

Frequency Down-conversion

Receiving filter

Sampling at TEqualizing

filter Detection

Coherent Detection – Binary Case Use a single correlator

Use two correlators

Coherent Detection – Decision

021

2

0

2

0

2

2

0

1

0

1

2)(

2

1exp

2

1)|(

2

1exp

2

1)|(

aaTz

azszp

azszp

Conditional probability functions

02: Noise variance

0: standard deviation

Coherent Detection – PSK Review

The information is contained in the instantaneous phase of the modulated carrier

For a binary PSK: 0 and 180 are used In fact, PSK can be views as ASK signal with the

bipolar carrier amplitudes The ideal detector requires perfect knowledge of

the un-modulated carrier phase at the receiver

Coherent Detection Binary PSK (BPSK)

TttT

Et

T

Ets

TttT

Ets

0,cos2

cos2

)(

0,cos2

)(

002

01

)()()(&)()()(

cos2

)(:function basis a Assuming

1121211111

01

tEtatstEtats

tT

t

EdtttntEEszE

EdtttntEEszE

tT

t

T

T

0 12

112

0 12

111

01

)()()(|

)()()(|

cos2

)( reference with thez(T) of valueexpected The

Take notes

Coherent Detection Sampled Matched Filter

Nyquist rate

Sampling time needs to be equal to or less than the symbol time

Shift to the register

Pay attention!

)3()1()( nsnNsnc iii

1

0

1

0

)()()(

)()()(

K

niii

N

nii

ncnkskzE

ncnkrkz

Coherent Detection Difference between MF and Correlators

If the timing of the MF and correlators are aligned, their outputs at the end of symbol time are identical

MF: A new output value is available in response to each new input sample; Equated to several correlators operating at different starting points of the input time series

Correlators: Computes an output once per symbol time

Timing errors in the correlator badly degraded performance

Example: Consider the waveform set s1(t)=At and s2(t)=-At,

t=[0,T] where k=0,1,2,3. Illustrate how a sampled matched filter can be used to detect a received signal from this sawtooth waveform set in the absence of noise.

Coherent Detection Multiple Phase-Shift Keying

M

it

T

Etsi

2cos

2)( 0

)(2

sin)(2

cos

)()()(

sin2

)(&cos2

)( :Choose

21

2211

0201

tM

iEt

M

iE

tatats

tT

ttT

t

iii

Coherent Detection – Multiple PSK

Inphase component

Quadrature component

Noise estimate of the transmitted

Probability of Bit Error for Coherently Detected BPSK

Detection rule:

otherwise)(

02

if)(

2

2101

ts

aa z(T)ts

Two types of errors:

0

0

)|()|(

)|()|(

22

11

r

r

dzszpseP

dzszpseP

Probability of Bit Error for Coherently Detected BPSK

)()|()()|( 2211 sPsePsPsePPB

A priori probability (equally likely)

dz

az

dzszpsePsePP

aar

aarB

2

0

2

2/0

2/ 221

2

1exp

2

1

)|()|()|(

210

210

00

212

2/

02

2

22exp

2

1

/)(

021 N

EQ

aaQdu

uP

azuLet

bu

aauB

Q function: complementary error function/co-error function

The standard deviation of the noise

2/ variancenoise the: 020 N

Probability of Bit Error for Binary Coherent Signals

In general, the BER becomes

vectorssignalbetween angle the:

signals obetween twt coefficienn correlatio-cross timeThe cos

)1(

2exp

2

1

0/)1(

2

0

N

EQdu

uP b

NEBb

Binary PSK: 1

Binary FSK: 02/

OOK: 02/

Example: Find the expected number of bit errors made in one day for a

BPSK system with a bit rate of 5000 bps. The received waveforms and are coherently detected with a matched filter. The value of A is 1mV. Assume that the single-sided noise power spectral density is

and that signal power and energy per bit are normalized relative to 1 ohm load.

tAts 01 cos)( tAts 02 cos)(

HzwN /10 110

Special Case – Detection

Case 1: signal energy is different

Case 2: the decision threshold is not at the middle point

0

0

)|(2

1)|(

2

121 r

r

B dzszpdzszpP

Example

NonCoherent Detection Differential PSK (DPSK)

The procedure of encoding the data differentially No attempt is made to determine the actual value of the

phase of the incoming signal

)()(cos2

)(

...2,1,0,)(cos2

)(

tnttT

Etr

MiTtttT

Ets

io

ioi

Typically assumed as a random variable uniformly distributed between zero and 2

NonCoherent Detection cont.

Detection of Differential PSK (DPSK) Matched filters are not possible – less efficient The carrier phase of the previous signaling interval can be

used as a phase reference The detector calculates the coordinate of the incoming

signal by correlating it with locally generated waveforms The detector then measures the angle between the currently

received signal vector and the previously received signal vector

NonCoherent Detection – Binary DPSK

addition 2-modulo :

)()1()(

)()1()(

kmkckc

kmkckc

Read your textbook

…0111011011…

… 01110110111

NonCoherent Detection – Binary DPSK

Optimal: requires a reference carrier in frequency but not necessarily in phase with the received carrier

Complex envelope: inphase component and quadrature component of the carrier wave

ttyttx

tjttjytxts

sin)(cos)(

)sincos)()(Re)(

Example The bit stream 1 0 1 0 1 0 1 1 1 1 is to be transmitted

using DPSK modulation. Show the encoded message (first bit is 1) and the detected message.

addition 2-modulo :

)()1()(

)()1()(

kmkckc

kmkckc1 0 1 0 1 0 1 1 1 1

1

1

0 0 1 1 0 0 1 0 1 0

1 0 0 1 1 0 0 0 0 0

Example When cables are installed in a building, it is not

unusual for the engineers to get the connections of the twisted pair reversed. How can a binary signaling scheme be designed to cope with this eventuality and maintain correct polarity data transfer?

Probability of Bit Error for Binary DPSK The decision is based on the phase difference

between successively received signals

TttT

Etx

TttT

Etx

0cos2

)(

0cos2

)(

02

01

Ttxxorxxts

Ttxxorxxts

20),(),()(

20),(),()(

12212

22111

Probability of Bit Error for Binary DPSK

0

exp2

1

N

EP b

B

Example A DPSK transmitter can generate an average power

of 1 nW at the input to a receiver which has a noise power density of 0.5  10-12 Watts/Hz. If the symbol rate is 100 symbols per second, what is the BER performance for a DPSK decoder in the receiver?

Let us work on bandwidth again… A binary PSK modem is designed to work within a

bandwidth of 8 kHz. What is the maximum data rate that can be delivered if a raised cosine filter with a = 1 is used?

Answer: 4kHz

FSK Generation FSK

The information contained in the frequency of the carrier

Insensitive to amplitude fluctuations in the channel

Generating FSK signals Switching between distinct frequency sources Voltage Controlled Oscillator Quadrature Modulator

MiTttT

Ets ii ,...2,1,0,cos

2)(

FSK Generation Switching between distinct frequency sources

FSK Generation Voltage Controlled Oscillator

FSK Generation Quadrature (Vector) Modulator

Symbol 1: c+

Symbol 1: c-

Summing appropriate amount of an in-phase and quadrature version of the carrier signal

tccos tcsin

Coherent FSK Detection

Decision: if the output from the mark filter is larger than that from the space filter, a decision is made that a mark signal was transmitted.

A matched filter demodulator is optimum because its filters are matched” to the transmitted signal so that their response to the desired signal is maximized with respect to their noise response.

NonCoherent Detection of FSK

Pass the signal through two bandpass filters turned to the two signal frequencies

Data can be recovered using an envelope detector [diode + smoothing filter]

Detect which has the larger output averaged over a symbol period

NonCoherent Detection of FSK

Phase-locked loop: A voltage controlled oscillator: output frequency is proportional to the input voltage A phase detector: produce a voltage output proportional to the phase difference A loop filter: control the dynamics of the feedback circuit

Advantages/Disadvantages of FSK Advantages

A constant envelope modulation insensitive to amplitude variation in the channel compatible with non-linear transceivers

Detection is based on relative frequency changes does not require absolute frequency accuracy in the channel

Disadvantages Less bandwidth efficiency BER performance is worth than of PSK

Probability of Error

Example What is the bit error probability for non-coherent

binary FSK for an Eb/N0 value of 10 dB? What approximate Eb/N0 is required to achieve the same BER performance of coherent FSK and PSK?

Example: page 240, 4.17 Consider that a BFSK domodulator/detector has s

synchronization error consisting of a time bias pT, where p is a fraction of the symbol time T. In other words, the detection of a symbol starts early (late) and concludes early (late) by an amount pT. Assume equally likely signaling and perfect frequency and phase synchronization. Find the general expression for bit-error probability as a function of p. If the received SNR is 9.6 dB and p=0.2, compute the value of

degraded BER due to the timing bias. If one did not compensate for the timing bias in this example, how

much additional SNR must be provided in order to restore the BER that exists when p=0.

Amplitude Shift Keying [not from the textbook]

ASK The information contained in the amplitude of the carrier

On-off Keying: the simplest form of bandpass data modulation

MiTttT

Ets i

i ,...2,1,0,cos2

)(

ASK Generation Linear modulator: an ASK signal can be realized

using a mixer to multiply the carrier with the baseband symbol stream

ASK Generation Switch

Binary ASK: switch to gate the carrier on and off, driven by the data signal.

M-ary signals: with differing amplitudes to represent the required number of symbol states

Non-linear process

ASK Generation Bandpass filtering method

Filter is needed after modulation A high frequency modulated data signal can be eliminated

ASK Generation Baseband filtering method

Using the mixer-based approach the baseband data stream can be pre-filtered using a low pass (root raised cosine) filter

Non-Coherent Detection Envelope detection: the information is conveyed in

the amplitude or envelope of the modulated carrier signal

A diode rectifier and smoothing filter

Coherent Detection By mixing the incoming data signal with a locally

generated carrier reference and selecting the difference component from the mixer output.

Coherent vs. Non-Coherent Phase or vector representation diagram

Case study: off state with noise Non-coherent: envelope detection N Coherent: N/2

Carrier Recovery Method 1: send a reference signal along with the

data signal Method 2: recover the carrier from the modulated

data signal Phase-locked loop (PLL)

By locking an oscillator to the phase of the incoming carrier when a carrier-on symbol is sent, and holding this oscillator phase when the carrier is off, it is possible to produce the required coherent reference.

Matched Filter Matched filter

Baseband transmission For optimizing the signal to noise ratio at the output of a

data receiver was discussed. Assumption: if the coherent detection is used

A matched filter pair such as the root raised cosine filters can thus be used to shape the source and received baseband data symbols in ASK

Timing Recovery Early-late gate synchronizer

The optimized filters are used Matched filter for instance

One with a slightly advanced timing reference One with a slightly retarded timing referenceComparator: periodically compared to see which is the larger The optimum timing signal is passed to a third data detector.

Example A coherent ASK demodulator has a 5o error in its

locally generated carrier reference. What will be the degradation in noise power immunity compared with an ideal demodulator?

M-ary Bandpass Modulation* Common knowledge

In principle, we can use any number of symbols for converting digital information.

A practical limit on the number of states to be used: the ability of receiving equipment to accurately resolve the individual states

A practical limit on the number of states to be used: the levels of noise and distortion introduced by the cannel and by the Tx and Rx units

Example: telephone modem 1024 symbol states vs. cellular systems two or four states

M-ary Bandpass Modulation M-ary signaling review

The processor considers k bits at a time The modulator produce one of M=2k waveforms

Does M-ary signaling improve the system performance? Error performance Bandwidth performance

M-ary ASK Implementation of M-ary ASK

Extension of binary ASK

Mixer at TX: to multiply the carrier with the baseband signal Coherent detectionMixer at RX: to multiply the received signal with a locally generated carrier referenceFilter: to select the DC value

M-ary ASK Performance

No opportunity to exploit orthogonally BER performance

Sensitivity to amplitude change Need for reasonable linearity

Constellation diagram* A representation of a signal modulated by a digital

modulation scheme To display the signal as a two-dimensional scatter

diagram in the complex Represents the possible symbols that may be selected

by a given modulation

M-ary FSK Increasing the noise immunity to achieve reliable

date transmission Possibility of using both orthogonal symbols or non-

orthogonal symbols Example: an orthogonal 8-ary FSK set with a symbol rate

of 1200 symbols/sec for coherent detection 1000Hz, 1600Hz, 2200Hz, 2800Hz, 3400Hz, 4000Hz, 4600Hz, and 5200Hz (same staring phase)

M-ary FSK Performance

M BER performance [cost: bandwidth]

M-ary PSK: Quadrature PSK PSK modulation scheme with four phase states

0, 90, 180, and 270 Twice the speed of BPSK in the same bandwidth

Modulator

Half the rate of the input data

Shape the data pulses in each channel

Continue…

M-ary PSK: Quadrature PSK

M-ary PSK: Quadrature PSK Demodulator

M-ary PSK: Quadrature PSK BER performance

Theoretical identical to that for BPSK

Differential QPSK Detection

/4 QPSK Widely used in the majority of digital radio modems Two identical constellations which are rotated by 45°

(π / 4 radians) with respect to one another reduces the phase-shifts from a maximum of 180°, but only to a maximum of 135° the filtered QPSK signal never passes through zero

Offset QPSK Staggering the input data streams to the two

quadrature BPSK modulators by half of symbol period

Symbol Error Performance

How about bit error probability? [Less]

How about bandwidth efficiency?

Probability of Symbol Error Equally-likely coherently detected M-ary PSK

Differentially coherent detection of M-ary DPSK

symbolper energy :log

sin2

2)(

2

0

MEE

MN

EQMP

bs

sE

MN

EQMP s

E2

sin2

2)(0

Probability of Symbol Error Equally likely coherently detected M-ary orthogonal

FSK

Equally likely noncoherently detected M-ary orthogonal FSK

0

)1()(N

EQMMP s

E

0

020

2exp

2

1)(

)!(!

!

exp)1(exp1

)(

N

EMMP

jMj

M

j

M

jN

E

j

M

N

E

MMP

sE

sM

j

jsE

Bit Error Rate vs. Symbol Error Rate

For orthogonal signals

For multiple phase signals

1

2/

12

2 1

M

M

P

Pk

k

E

B

M

PP E

B2log

Example: page 240, 4.12

Consider a 16-ary PSK system with symbol error probability 10-5. A Gray code (binary) is used for the symbol to bit assignment. What is the approximate bit error probability?

For a 16-ary orthogonal FSK system.

654

31

65

103.51012

2

12

2

105.24

10

Ek

k

B

EB

PP

k

PP

Gray code A binary numbering system two successive values differ in only one digit

Example: 00 01 11 10 Example: 000 001 011 010 110 111 101 100

Applications Sensor: angle detection Digital communication: error detection