1. ECEN5533 Modern Commo Theory Dr. George Scheets Lesson #25 11 November 2014 n Read Section 6.4 – 6.5 n Problems: 9.2, 9.4, 9.7, 6.2, 6.3 n Design #2 due 11 November n Reworked Exam #2 due 18 November

2. ECEN5533 Modern Commo Theory Dr. George Scheets Lesson #26 13 November 2014 n Read Section 6.6 – 6.9, 7.1 n Problems: 5.8, 6.4 & 5, 9.8 & 9 n Reworked Exam #2 due 18 November n Final Exam, 0800 – 0950, Tuesday, 8 December

3. ECEN5533 Modern Commo Theory Dr. George Scheets Lesson #27 18 November 2014 n Read Section 7.2 - 7.3 n Problems: 6.9, 6.11, 6.13, 7.1 n Reworked Exam #2 due today n Reworked Design #2 due 25 November n Final Exam, 0800 – 0950, Tuesday, 9 December

4. ECEN5533 Modern Commo Theory Dr. George Scheets Lesson #28 20 November 2014  Read Section 7.4  Problems 7.3, 7, 10, & 12 n Reworked Design #2 due 25 November n Late Fee -1 per working day n Final Exam, 0800 – 0950, Tuesday, 8 December

5. ECEN5533 Modern Commo Theory Dr. George Scheets Lesson #29 25 November 2014  Read Section 12.1  Problems 7.16, 12.3 & 12.4 n Reworked Design #2 due today n Comprehensive Final Exam (No Rework) u Tuesday, 9 December, 0800-0950

6. ECEN5533 Modern Commo Theory Dr. George Scheets Lesson #30 2 December 2014  Problems 12.9, 10, 11, & 21  Comprehensive Final Exam (No Rework)  Tuesday, 9 December, 0800-0950

7. ECEN5533 Modern Commo Theory Dr. George Scheets Lesson #31 4 December 2014  Radar Set & Old Finals  Comprehensive Final Exam (No Rework)  Tuesday, 9 December, 0800-0950

8. Design #2: Digital Satellite RFP n Lowest Working Bid $20.54 Million u Matt Gaalswyck u Promoted to Senior Engineer II at MegaMoron n ASK @ 2.2 GHz n 84-1 Compression, (2.54,1) FEC n Tower at city center u G1 = 2.493 over 285 degrees u G2 = 1.116 over 75 degrees, aimed due N n 1,000,000 receiver antenna elements n 36.13 dB margin n Largest Cost: Geek Telecom @ $7.75 M

9. Point Spreads of Completed Stuff n Quiz #1 (20 points) Hi = 19.1, Low = 9.7, Ave = 13.45, σ = 3.20 n Quiz #2 (20 points)* Hi = 15.9, Low = 10.6, Ave = 13.52, σ = 2.38 n Exam #1 (100 points)* Hi = 86, Low = 46, Ave = 65.83, σ = 17.23 A > 85, B > 69, C > 59, D > 49 n Exam #2 (100 points)* Hi = 97, Low = 31, Ave = 70.33, σ = 24.33 A > 90, B > 78, C > 68, D > 58 n Design #1 (70 points) Hi = 66, Low = 59, Ave = 65.80, σ = 3.39

10. Anything in Circles is Fair Game on Final Exam S Read HW Notes

11. Bit Error Rate Unsatisfactory? n System designer has several options: u Use FEC codes u Increase received signal power u Use more effective modulation technique F Best for baseband: + & - square pulses F Best for RF Binary system? PSK M-Ary system? QPSK or QAM u Slow down the transmitted message symbol rate u Decrease receiver T system

12. FEC Examples n Matched Filter Detector (MFD) n MFD P(BE) gets worse as bit rate increases u h(t) = 1; 0 < t < T, for an integrator u H(f) = sinc with a phase shift u Integration time becomes shorter u H(f) becomes wider, less of a low pass filter

13. FEC Examples n In the limit, as bit interval T approaches zero u # of independent samples approaches 1 u MFD P(BE) approaches SSD P(BE) n Suppose you have a system where u P(BE) = 0.02 for MFD at bit rate R (no FEC) u P(BE) = 0.03 for MFD at bit rate 2R (2:1 FEC) u P(BE) = 0.04 for MFD at bit rate 3R (3:1 FEC)

14. Matched Filter Detector & No coding: Block Diagram Source Channel Channel Coder Symbol Detector: Matched Filter P(Data Bit Error) =.02

15. MFD 2:1 FEC Source Source Coder: Input = 1 bit. Output = Input + Parity bit. Channel Channel Coder Symbol Detector: Matched Filter Source Decoder: Looks at blocks of 2 bits. Outputs 1 bit. R application bps 2R code bps 2R code bps R app. bps P(code bit error) =.03

16. Example) MFD 2 bit code words n Suppose you transmit each bit twice, smaller bit width will cause P(Code Bit Error) to increase to, say 0.03 u Legal Transmitted code words; 00, 11 u Possible received code words 00, 11 (appears legal, 0 or 2 bits in error) 01, 10 (clearly illegal, 1 bit in error) P(No code bits in error) =.97*.97 =.9409 P(One code bit in error) = 2*.97*.03 =.0582 P(Both code bits in error) =.03*.03 =.0009 u Decoder takes 2 code bits at a time and outputs 1 data bit If illegal code word received, it can guess 0 or 1. 94.09% + 5.82%(1/2) = 97% of time correct bit output.09% + 5.82%(1/2) = 3% of time the incorrect bit is output u FEC makes it worse: 3% data bit error vs 2% No Coding

17. MFD 2:1 FEC Source Source Coder: Input = 1 bit. Output = Input + Parity bit. Channel Channel Coder Symbol Detector: Matched Filter Source Decoder: Looks at blocks of 2 bits. Outputs 1 bit. R application bps 2R code bps 2R code bps R app. bps P(code bit error) =.03 P(data bit error) =.03 P(Data Bit Error) =.02 when FEC not used.

18. Typical FEC Performance Eb/NoEb/No P(BE) Coded Plot changes as type of symbol, type of detector, and type of FEC coder change. Uncoded Plot changes as type of symbol, and type of detector change. Last example is operating here. There generally always is a cross-over point. The max possible P(BE) = 1/2.

19. MFD 3:1 FEC Source Source Coder: Input = 1 bit. Output = Input + two parity bits. Channel Channel Coder Source Decoder: Looks at blocks of 3 bits. Outputs 1 bit. Symbol Detector: Matched Filter P(code bit error) =.04 R application bps 3R code bps 3R code bps R app. bps

20. Example) MFD 3 bit code words n Transmit each bit thrice, P(Bit Error) again increases to, say 0.04, due to further increase in the bit rate. u Legal Transmitted code words; 000, 111 u Possible received code words 000, 111 (appears legal, 0 or 3 bits in error) 001, 010, 100 (clearly illegal, 1 or 2 code bits in error) 011, 101, 110 (clearly illegal, 1 or 2 code bits in error) P(No code bits in error) =.96*.96*.96 =.884736 P(One code bit in error) = 3*.96 2 *.04 =.110592 P(Two code bits in error) = 3*.96*.04 2 =.004608 P(Three code bits in error) =.04*.04*.04 =.000064 u Decoder takes 3 bits at a time & outputs 1 bit. Majority Rules. 88.4736% + 11.0592% = 99.5328% of time correct bit is output.0064% +.4608% = 0.4672% of time incorrect bit is output u FEC makes Data BER better (.5% vs 2%) @ thrice the bit rate

21. MFD 3:1 FEC Source Source Coder: Input = 1 bit. Output = Input + two parity bits. Channel Channel Coder Source Decoder: Looks at blocks of 3 bits. Outputs 1 bit. Symbol Detector: Matched Filter P(code bit error) =.04 R application bps 3R code bps 3R code bps R app. bps P(data bit error) =.005 P(Data Bit Error) =.02 when FEC not used.

22. Rate 1/2 Turbo Coder n u k (data) & v k (parity bits) are transmitted to far side n v k = v 1k 1/2 of the time & v 2k other half Source: Figure 8.26 from Sklar's Digital Communications

23. Rate 1/2 Turbo Decoder n x k (corrupted data) & y k (corrupted parity bits) n y k = y 1k 1/2 of the time & y 2k other half Source: Figure 8.27 from Sklar's Digital Communications ← Matched Filter ← Matched Filter

24. Rate 1/2 Turbo Coding Performance n P(data bit error) = 0.00001 when Eb/No = 0.2 dB & 18 reps n P(bit error) = 0.07395 for BPSK when Eb/No = 0.2 dB Source: Figure 8.28 from Sklar's Digital Communications

25. Performance n Uncoded BPSK n Hard coded Block or Convolutional BPSK n Soft coded Convolutional BPSK u 2 dB increase in effective Eb/No Compared to Hard Convolutional Decoding n Turbo Coded BPSK u Big time increase in effective Eb/No u Can get you close to Shannon Limit n All of above require an increase in the bit rate u Need more bandwidth, or go M-Ary n Trellis Coded Modulation u 3 db – 6 dB increase in effective Eb/No u Doesn't require an increase in bit rate Improved Performance

26. Low Density Parity Check Codes n Developed by Robert Gallagher, 1963 MIT Grad u Low Density → Few 1's in rows & columns of H u Impractical to implement then n Linear Block Codes u Huge block sizes u DVB-S2 (Video) Code 43,200 data bits & 21,600 parity bits n Offers comparable performance to Turbo Codes u Being used in some of the newest standards F 10 Gbps Ethernet over twisted pair F 802.11n & 802.11ac (optional) u Turbo Codes: Easy to Encode, Hard to Decode LDPC Codes: Hard to Encode, Easy to Decode

27. Spread Spectrum n Two Kinds u Direct Sequence u Frequency Hopping n Advantages u Interference Suppression u Low Probability of Exploitation u Multipath Effects are Reduced u Code Division Multiple Access n Uses u 3G Cell Phone Standards u Lower to mid speed WiFi (IEEE 802.11)

28. +1 time +1 +1 Traffic (9 Kbps) Spreading Signal 27 Kcps Transmitted Signal 27 Kcps +1 DSSS - Transmit Side

29. Wireless X 27 Kcps Square Pulses cos(2πf c t) BPSK output 27 Kcps 90% of power in 54 KHz BW centered at f c Hertz X cos(2πf c t) BPSK input 27 Kcps + noise 27 Kcps Square Pulses + filtered noise RCVR Front End RF Transmitter Low Pass Filter

30. time +1 +1 Despreading Signal 27 Kcps Received Signal 27 Kcps +1 +1 time Recovered Traffic 9 Kbps DSSS-Receiver

31. DSSS Receiver n If the proper source is transmitting... n...and the receiver has the correct despread sequence... n...and the sequence is properly synchronized... n...the original message is recovered.

32. DSSS Receiver n If another source is transmitting... n...the receiver will have the wrong despread sequence... n...and the output will be garbage.

33. time+1 Received Signal #2 27 Kcps +1 time Recovered Garbage from 2nd signal +1 time +1 Despreading Signal 27 Kcps +1 DSSS-Receiver

34. Recovered Garbage at 2nd receiver Receiver Matched Filter Detector Message Output is a random sequence of 0’s & 1’s +1 time +1

35. DSSS Receiver n If both sources are transmitting... n...the bit detector will be fed the sum of the results.

36. Input to Matched Filter Detector (sum) +1 time Recovered Traffic 9 Kbps time Recovered Garbage from 2nd signal +1 +2+1 +1-2 time DSSS-Receiver +2

37. Receiver Matched Filter Detector Output Additional signals transmitting at the same time increase the apparent noise seen by our system. Message BER will increase. +1 time Input to Matched Filter Detector (sum) +2-2 time T Bit +2

38. FDM FDMA WDM frequency time Different channels use some of the bandwidth all of the time. 12345

39. TDM TDMA frequency time Different channels use all of the bandwidth some of the time. Predictable time assignments. 1 2 3 1 etc.

40. CDMA frequency time Different channels use all of the bandwidth all of the time. Channels use different codes. Other channels cause noise-like interference.

41. CDMA: 3D View code #1 code #2 code #3 frequency time

42. CDMA vs FDMA n Example) Given 10 MHz Channel & Coding Gain of 1,000 u CDMA will support 75 users u FDMA will support 900 users u All things being equal... Power Out, Path Loss, Antenna Gains, etc. n In real world, all things aren't always equal

43. CDMA vs FDMA (or TDMA) n Narrowband Noise u May knock out some FDMA or TDMA channels n Severe Multi-path Environment u May knock out some FDMA channels n Easier to add users u Transmit with different code (CDMA) u Must find empty time slot or frequency band n Easier to use Variable Rate Coder u Voice Coder with Silence Suppression F Doubles potential capacity

44. DSSS Wireless Example X 100 Kcps Zero Mean Square Pulses cos(2πf c t) BPSK output 100 Kcps 90% of power in 200 KHz BW centered at f c Hertz X cos(2πf c t) BPSK input 100 Kcps + noise + 2nd DSSS Signal 100 Kcps Zero Mean Square Pulses + filtered noise RCVR Front End RF Transmitter Low Pass Filter (Wide Band) Band Pass Filter (Wide Band)

45. Spread Spectrum Receiver X cos(2πf c t) BPSK input 100 Kcps + noise + 2nd DSSS Signal 100 Kcps Zero Mean Square Pulses + filtered noise + 2nd DSSS signal RCVR Front End Low Pass Filter (Wide Band) Band Pass Filter (Wide Band) X Despread Sequence Low Pass Filter (Narrow Band) 10 Kbps Message + noise + 2nd DSSS interference (noise like)

46. Radar XMTR RCVR Switch Antenna Same antenna normally used. Either Transmitter or Receiver connected at any time.

47. F-15 Eagle RCS ≈ Barn Door?

48. F-117 Nighthawk RCS ≈ Hummingbird = 0.025 m 2 ?

49. B-2 Spirit RCS ≈ 0.1 m 2 ?

50. Impulse Response Non-stealth vs Stealth Source: Cheville & Grischkowsky, "Time Domain THz Impulse Response Studies", Applied Physics Letters, October 1995

51. Impulse Response Looking Down from Top Left Source: Cheville & Grischkowsky, "Time Domain THz Impulse Response Studies", Applied Physics Letters, October 1995

52. Voyager II http://voyager.jpl.nasa.gov/index.html nLnLnLnLaunch uAuAuAuAugust 1977 nJnJnJnJupiter fly-by uJuJuJuJuly 1979 nSnSnSnSaturn fly-by uAuAuAuAugust 1981 nUnUnUnUranus fly-by uJuJuJuJanuary 1986 nNnNnNnNeptune fly-by uAuAuAuAugust 1989 n1n1n1n15.93 Billion Km u1u1u1u106.5 AU uNuNuNuNovember 2014 Source: JPL source: http://voyager.jpl.nasa.gov/

53. Voyager Spacecraft source: September 1990 IEEE Communications Magazine

54. NASA Deep Space Network 70 m diameter parabolic source: http://deepspace.jpl.nasa.gov

55. Voyager FEC Coding source: Science, Summer 1990

56. BER Performance at Jupiter source: Science, Summer 1990 CODING NO CODE Target BER: Imaging 5(10 -3 ) Non-Imaging: 5(10 -5 ) Command: 1(10 -5 )

57. BER Performance at Saturn source: Science, Summer 1990 CODING Same System Configuration as at Jupiter Target BER: Imaging 5(10 -3 ) Non-Imaging: 5(10 -5 ) Command: 1(10 -5 ) CODING Slowed bit rate compared to Jupiter

58. BER Performance at Uranus source: Science, Summer 1990 Target BER: Imaging 5(10 -3 ) Non-Imaging: 5(10 -5 ) Command: 1(10 -5 ) CODING Reduced R Increased Aer Decreased Tsys compared to Saturn CODING Same System Configuration as at Saturn

59. Signal * Wideband Noise

60. BER Performance at Neptune source: Science, Summer 1990 Target BER: Imaging 5(10 -3 ) Non-Imaging: 5(10 -5 ) Command: 1(10 -5 ) CODING Reduced R Increased Aer Rebuilt antennas Additional coupling CODING Same System Configuration as at Uranus

61. NRAO's Very Large Array image source: http://www.vla.nrao.edu/

62. MIMO n Used in latest Cell & Wireless LAN protocols u WiFi 802.11n & 802.11ac u LTE 4G Cellular n Potential Benefits u Steerable Beams F Increased antenna gain u Spatial Multiplexing F Transmit several signals over (ideally) independent paths F Increase usable BW u Spatial Diversity F Several Versions of XMTR signal received F Improves BER

63. MIMO antenna source: http://www.pcmag.com/article2/0,1759,1822020,00.asp Belkin Wireless Pre-N Router F5D8230-4

64. MIMO Example fc = 300 MHz λ = 1 meter Same signal fed to both antennas. Beam shoots out both sides at 90 degree angle. λ/2 Directivity Strength

65. MIMO Example fc = 300 MHz λ = 1 meter Signal to left antenna advanced by 333.3 picosecond ( = 10% wavelength) with respect to right antenna. λ/2 Directivity Strength

66. MIMO Example fc = 300 MHz λ = 1 meter Signal to left antenna delayed by 333.3 picosecond ( = 10% wavelength) with respect to right antenna. λ/2 Directivity Strength

67. MIMO Example fc = 300 MHz λ = 1 meter Signal to left antenna delayed by 833.3 picosecond ( = 25% wavelength) with respect to right antenna. λ/2 Directivity Strength

68. MIMO Example fc = 300 MHz λ = 1 meter Signal to left antenna delayed by 1 2/3 nanosecond ( = 50% wavelength) with respect to right antenna. λ/2 Directivity Strength

69. Directivity:.4λ spacing source: www.orbanmicrowave.com/The_Basics_of_Antenna_Arrays.pdf # elements red = 2 green = 5 blue = 10

70. Directivity: 5 elements source: www.orbanmicrowave.com/The_Basics_of_Antenna_Arrays.pdf spacing red =.2λ green =.3λ blue =.5λ

71. SISO n Potential Benefits u Can use Steerable Beams F Increased antenna gain u Spatial Diversity F Several Versions of XMTR signal received F Improves BER u Spatial Multiplexing F Transmit several signals over independent paths F Increase usable BW

72. MISO n Potential Benefits u Transmitter can use Steerable Beams F Increased antenna gain u Spatial Diversity F Several Versions of XMTR signal received F Improves BER u Spatial Multiplexing F Transmit several signals over independent paths F Only one signal per Single Antenna (SO) Receiver F Increase usable BW

73. Needed for Spatial Multiplexing: n Accurate Knowledge of RF channel n Can get by… Periodically transmit known sequences u x1(t) = string of logic 1's u x2(t) = alternating 1's and 0's n Look at relative strength at two outputs u Baseband x1(t) = constant value u Baseband x2(t) = peak-to-peak value

74. SIMO n Potential Benefits u Receiver can use Steerable Beams F Increased antenna gain u Spatial Diversity F Several Versions of same XMTR signal received F Feed signal with strongest power to MFD & FEC F Improves BER u Spatial Multiplexing F Receive several signals over multiple paths F Increase usable BW MF Detector Power = ? switch

75. Wish to Probe Further? n See… Multiple Antenna Techniques for Wireless Communications What Will 5G Be? What Will 5G Be? n (Links on 5533 Home Page)

76. OFDM n Resistant to narrow band interference u Which might knock out single carrier system n Resistant to multi-path n Spread Spectrum has same benefits u Requires extra BW compared to OFDM n Usually implemented with FFT & IFFT n Used in latest Wireless Protocols u WiFi 802.11n & 802.11ac u LTE 4G Cellular

77. Standard Single Carrier Modulation frequency time Message power is multiplied by a carrier with a single center freq. M-ASK, M-PSK, M-QAM Channel 1 Example: 8 Mbps bit stream carried by B-PSK with 16 MHz null-to-null BW

78. Orthogonal FDM frequency time Channels split into sub-channels Bits parceled out to sub-channels Advantage: Sub-channel bit rates can be modified to cope with interference Less susceptible to multipath Channel 1

79. FDM with Multi-path XMTR RCVR direct path bounce path direct path pulses bounce path pulses Signal sum seen by Receiver T1T2T3 Symbol decision intervals at Receiver. The third bit is obliterated by multi-path. T3 time delay

80. OFDM with Multi-path direct T3 bounce direct bounce direct bounce T2T1 Matched filter detector will work OK. delay Slower symbol rate over each subchannel.