EE 400: Communication Networks (101)
Dr. Wajih A. Abu-Al-Saud
Latestupdateofthislecturewason29112010
Ref: A. Leon Garcia and I. Widjaja, Communication Networks, 2nd Ed. McGraw Hill, 2006
Lecture21:DigitalTransmissionFundamentals Basebandvs.PassbandDigitalModulation
The line codes described above produce baseband signals. Such baseband signals are suitable for transmissionover cables (which represent baseband channels). When attemptingtotransmit a digital signal over a passband channel (a channel that allows frequencies from some frequency f1 > 0 to another frequency f2 > f1 to pass andstops other frequencies), a modified process ofthe line coding describedinthepreviouslecturebecomesnecessary.Theprocessisdigitalmodulation.Tounderstand theprocessofdigitalmodulation,wehavetodiscusstheprocessofgeometricrepresentationofdigital modulations. The difference betweenbaseband and passband digital transmissionthat weare interested in here is that the difference between the different pulses that are transmitted in baseband digital communication appears in the amplitude of the pulses, while the difference between the different pulses that are transmitted in passband digital communication appears in either the amplitude, the phaseorthefrequencyofthecarriersusedtocarrytheinformation.
GeometricRepresentationofDigitalModulations
Tounderstandtheprocessofgeometricrepresentationofdigitalmodulations,letusfirstdiscussthe reasonsfordoingthisprocess.
Importance of Geometric Representation of Digital Modulations
Theprocessinwhichdigitalmodulationisusedinvolvesthefollowingsimpleprocesses: 1. The digital data is received from the data source (for example, a computer that needs to transmittext,oradigitalcamerathatneedstotransmitapicture) 2. Thebitsaredividedintogroupsofnbitseach,thevalueofnis1forbinarydigitalmodulation techniques,and n = log 2 M bits(suchthat M = 2 n )inMarymodulationtechniques 3. Foreachgroupofbits,onepulseoutofasetof M pulsesisselectedthatcorrespondstothe combinationofbitsineachgroup 4. Thepulsesaretransmittedonepulseaftertheother 5. Thepulsesarereceivedoneaftertheotheratthereceiverintheformofacontinuoussignal 6. Thepulsesareseparatedfromeachother 7. Each pulse is detected (meaning that the receiver tries to determine which pulse was transmitted) 8. Thedetectedpulsesareconvertedtothecorresponding n bits 9. The n bits of the different pulses are recombined into their original format (bytes for ASCII characters,3bytesperpixelforimages,) 1
EE 400: Communication Networks (101)
Dr. Wajih A. Abu-Al-Saud
Latestupdateofthislecturewason29112010
Ref: A. Leon Garcia and I. Widjaja, Communication Networks, 2nd Ed. McGraw Hill, 2006
Lecture21:DigitalTransmissionFundamentals
Theaboveprocessisillustratedinthefollowingfigure:
EE 400: Communication Networks (101)
Dr. Wajih A. Abu-Al-Saud
Latestupdateofthislecturewason29112010
Ref: A. Leon Garcia and I. Widjaja, Communication Networks, 2nd Ed. McGraw Hill, 2006
Lecture21:DigitalTransmissionFundamentals
Otherstepsmaybeaddedtotheaboveindifferentdigitalcommunicationsystemsasneededincluding Coding/decoding (adding some redundancy bits to the transmitted data and using these redundancybitstodetect/correcterrorsthatoccurinthereceivedbits) Equalizationofthereceivedsignal(toremovethedistortingeffectsofthechannel)
Performingeachoftheabove9stepsissimpleexceptforstep7.Thewayofperformingtheprocessof step 7 (detecting which pulses were transmitted) may not be clear especially if we know that the receivedpulsesmaybeverydifferentfromthetransmittedpulsesasaresultofchanneldistortionand mostlychannelnoisethatwasaddedtothesignalasitwastravelinginthechannel.Thismeans: Ifwereceiveapulsethatisverysimilartooneofthetransmittedpulses,wecanfairlyassume thatthecorrespondingpulsewastransmitted, Ifwereceiveapulsethatisdifferentfromalltransmittedpulsesbecauseofnoise,wehaveto findawaytodeterminewhichpulseisthemostlikelyonethatwastransmitted?
Therefore, to perform step 7 we need a method for quantifying the similarity between the pulses (finding a quantity that indicates the similarity between the received pulse and each of the original noiselesspulses)fordeterminingwhichoftheseoriginalpulsesisclosesttothereceivedpulse.Todo this,weneedtorepresentsignalsgeometrically.
Concept
The concept of geometric representation of digital modulated signals requires the understanding of orthogonalsignalsororthogonalpulses.Observethefollowing: Apulse(whichisasignalthatisdefinedoveraspecificperiodoftime TS )maybeamultipleof anotherpulseifwecanrepresentoneofthemasarealvaluetimestheother.Example:
x1 (t ) = cos ( 2 f 0t ) x2 (t ) = 5cos ( 2 f 0t )
Clearly,
0 t TS 0 t TS
x 2 (t ) = 5 x 1 (t )
0 t TS
Apulsemaybeorthogonaltoanotherpulseiftheintegrationoftheproductofthetwopulses overthetimeintervaltheyaredefinedoverisequaltozero.Example:
x1 (t ) = cos ( 2 f 0t ) x2 (t ) = sin ( 2 f 0t )
0 t TS 0 t TS
3
EE 400: Communication Networks (101)
Dr. Wajih A. Abu-Al-Saud
Latestupdateofthislecturewason29112010
Ref: A. Leon Garcia and I. Widjaja, Communication Networks, 2nd Ed. McGraw Hill, 2006
Lecture21:DigitalTransmissionFundamentals
where TS is an integer multiple of
1 . If we multiply these two signals and integrate the f0
productovertheperiod 0 t TS wegetzero.Youcaneasilyverifythat
TS
x1 (t ) x2 (t ) dt = 0
Apulsemayneitherbeamultipleofsomeotherpulseororthogonaltoit.Thismeansthatthe projectionofthefirstpulseonthesecondisnonzero.Example:
x1 (t ) = cos ( 2 f 0t )
0 t TS 0 t TS
x2 (t ) = cos 2 f 0t + 4
Where TS isanintegermultipleof
TS 0
1 .Inthiscase,itisclearthat f0
x1 (t ) x2 (t ) dt 0 ,and
x2 (t ) c x2 (t )
for any real constant c
Giventheabovepoints,itisclearthatifwerepresentapulsebyavectorinaCartesianspacesuchthat: a. Thelengthofthevectorisproportionaltotheamplitudeofthepulse b. Theangleofthevectorisrelatedtothephaseofthecarrierofthepulse, then: 1. Two pulses x1 (t ) and x2 (t ) such that x2 (t ) = c x2 (t ) can be represented in the Cartesian spaceastwovectorsinthesamedirectionif c ispositiveoroppositedirectionif c isnegative such that the length of the vector x2 corresponding to x2 (t ) is c times the length of x1 correspondingto x1 (t ) . 2. Twopulses x1 (t ) and x2 (t ) suchthat x1 (t ) isorthogonalto x2 (t ) canberepresentedinthe Cartesianspaceastwoperpendicularvectors x1 and x2 withmagnitudesproportionaltothe magnitudesofthetwopulses. 3. Two pulses x1 (t ) and x2 (t ) such that x2 (t ) c x2 (t ) for any real constant c (i.e. the two pulses are not multiples of each other) and x1 (t ) x2 (t ) dt 0 (i.e. the two signals are not
0 TS
EE 400: Communication Networks (101)
Dr. Wajih A. Abu-Al-Saud
Latestupdateofthislecturewason29112010
Ref: A. Leon Garcia and I. Widjaja, Communication Networks, 2nd Ed. McGraw Hill, 2006
Lecture21:DigitalTransmissionFundamentals
orthogonaltoeachother)canberepresentedintheCartesianspaceastwovectors x1 and x2 thatareneitherinthesamedirectionnorperpendiculartoeachother.
Example 1
Considerthetwopulsesshownbelow
x 1 (t ) = cos ( 2 f 0t ) x 2 (t ) = 4cos ( 2 f 0t )
0 t TS 0 t TS
IfthefirstpulseisrepresentedbytheREDdotintheCartesianspacebelow,showtherepresentationof theotherpulse
Solution
Clearly,thesecondpulseisarealmultipleofthefirstone,or
x 2 (t ) = 4 x 1 (t )
0 t TS
So,thesecondpulseisrepresentedbyavectorthatisinthesamedirectionofthefirstonebutwitha magnitudethatis4timesasmuch.
Example 2
Considerthetwopulsesshownbelow
EE 400: Communication Networks (101)
Dr. Wajih A. Abu-Al-Saud
Latestupdateofthislecturewason29112010
Ref: A. Leon Garcia and I. Widjaja, Communication Networks, 2nd Ed. McGraw Hill, 2006
x 1 (t ) = cos ( 2 f 0t )
Lecture21:DigitalTransmissionFundamentals
0 t TS 0 t TS
x 2 (t ) = 3sin ( 2 f 0t )
IfthefirstpulseisrepresentedbytheREDdotintheCartesianspacebelow,showtherepresentationof theotherpulse
Solution
Clearly, the two pulses are orthogonal (if the first one has a phase of 0, the second one will have a phaseof90.Also,themagnitudeofthesecondoneis3timesthemagnitudeofthefirst.
Example 3
Considerthethreepulsesshownbelow
EE 400: Communication Networks (101)
Dr. Wajih A. Abu-Al-Saud
Latestupdateofthislecturewason29112010
Ref: A. Leon Garcia and I. Widjaja, Communication Networks, 2nd Ed. McGraw Hill, 2006
x 1 (t ) = cos ( 2 f 0t )
Lecture21:DigitalTransmissionFundamentals
0 t TS 0 t TS 0 t TS
x 2 (t ) = 3cos ( 2 f 0t + 60 ) x 3 (t ) = 2 cos ( 2 f 0t + 30 )
IfthefirstpulseisrepresentedbytheREDdotintheCartesianspacebelow,showtherepresentationof theotherpulses
Solution
Clearly,nopairofthethreepulsesareeithermultipleofeachotherorperpendiculartoeachother.We getthefollowing:
