수업일지 - 디지털통신(00-2학기)

수업일지 -- 디지털 통신 (3반) 2000학년도 2학기, 전기전공

여러분이 이 페이지를 읽는 날짜를 기준으로 하여

이전의 기록은 이미 수업시간에 다루었던 내용을 간추린 것이고 이후의 기록은 대략적인 계획입니다.

자주 들러서 예습/복습에 도움이 되기를 바라며, 열심히 공부하기 바랍니다.

담당교수: 송홍엽

FIRST QUARTER (GO TO 1st 2nd 3rd 4th quarter; top today bottom )

9월 2일(금78) 개강.

수업 계획서 배포
Chapter 1 강의자료 ohp자료 복사실에 - 30 pages
교재소개, class rule 설명
출석부확인
디지털 통신 모뎀 블록도(그림 1) 설명
random variables에 대한 간단한 quiz

9월 6일(수8)

Chapter 1 강의노트 + 논문 2편 복사실에 - 20 pages
공지사항 remind, 특히, 사진 보낼 것
source coding, channel coding의 개념.

9월 8일(금78)

digital modem의 개념
chapter 1 마무리 - signals and systems
- energy signal/power signal
- energy spectral density/power spectral density
- random variables/random processes
Signals and Systems 복습할 것 당부
Random Variables and Random Processes 복습할 것 당부 (연휴기간중)

9월 13일(수8) 추석연휴기간

9월 15일(금78)

Sections 2.1 - 2.7 : Formatting
- textual data, analog signal
- sampling theorem, S/N of quantizer
Sec. 2.8 Baseband Transmission
- PCM coding, Analysis and Synthesis coding
- NRZ, RZ, Phase coding, Multilevel binary
- dc component, self-clocking, error-detection, bandwidth compression, etc
HW#1: Chapter 2 Exercise Problems #2-#10, 마감: 다음 주 수요일 8교시 수업시간 시작시까지
강의노트 sections 2.1-2.8 복사실에.

9월 20일(수8)

Section 2.9
Basics of Signal space
Basic block diagram of transmitter/channel/receiver
receiver=demodulator+detector

9월 20일(수10,11) 보강

Section 2.9 continued
ONE BIG EXAMPLE of binary antipodal signaling
- transmit waveforms :1-dimensional
- channel model: simple additive, AWGN
- demodulation is to extract a number(test statistic)
- detection is to make a decision based only on the test statistic
- MAP decision
- ML decision
- decision region
- BER calculation
- Q-function

HW#2: Chapter 2 Exercise Problems #13-#22, 마감: 다음 주 금요일 7교시 시작시까지
HW#1 copy한 사람은 다시 풀어서 제출하시오. (스스로)

9월 22일(금78) 연습시간

Sections 2.9.2-2.9.4
HW#1 질문받기

9월 27일(수8)

Section 2.9 지난시간 내용 복습
remainings of Section 2.9
- theorem of irrelevance
- 일반적인 1-dim binary signaling system의 detector 구조 및 BER
- Unipolar signaling과 Bipolar signaling의 비교
Section 2.10
- 2-level VS 8-level PCM의 비교
- 평균 비트에너지, 평균 펄스에너지, data rate, bandwidth 의 비교

9월 27일(수10) 연습시간

Sections 2.11-2.12
강의노트 (Sections 2.9 - 2.11) 복사실에

9월 29일(금78)

연고전 휴강 -- 열심히 응원하세요. (잠실에서 출석을 부를까....)
HW#2 마감을 하루 앞당깁니다. 마감은 9월 28일(목) 5시까지. B613호 조교에게.

SECOND QUARTER (GO TO 1st 2nd 3rd 4th quarter; top today bottom )

10월 4일(수8) Chapter 3

Sections 3.1-3.2
Two primary causes for signal distortion (1) filtering effects (2) noise
Concept of signal space (revisited)
Vector space, norm, distance, inner product
Why we use signal space ?
1. Complexity of signal representation reduces dramatically while preserving properties of signals
2. Essential contribution of noise to the operation of receiver can be identified
Gram-Schimidt Orthogonalization Procedure : self-exercise

10월 4일(수10,11) QUIZ#1 -- Chapters 1 and 2

10월 6일(금78)

definition of correlation matrix of a signal set
classification: orthogonal, antipodal, bi-orthogonal, simplex signals
general problem of signal design: find M signals in N dimensional space such that the maximum (or, average) correlation is minimized
HW#3 Signal Design and Constellation Problems: Due 10월8일(수) 수업시간.
introduction to bandpass modulation
- amplitude, phase, frequency
- coherent vs noncoherent demodulation
- PSK, FSK, ASK, APK, etc

10월 11일(수8)

휴강
연습시간을 8교시에

10월 13일(금78)

Decision Region and Correlation Receiver

10월 18일(수8)

Coherent Detection of Binary Signals in AWGN (Chapter 3, part 2)
- 0. Model
- 1. Assumptions
- 2. Signal space representation
- 3. At the receiver
- 4. Statistics of n
- 5. Statistics of r
- 6. ML decision is Minimum distance decision
- 7. Optimum receiver
- 8. Further simplification
HW#4 Joint Gaussian random variables and transformations: Due 11월 1일 (수) 수업시간

10월 20일(금78)

8. Still further simplification
9. Performance
10. Various signals for (coherent) binary systems: BPSK, BFSK
HW#4: 3,4,5,6번 문항에서 sigma_z = sigma_y = sigma로 할것.

보충설명
- difference between MF and Correlator implementations
- difference between correlating with signals and correlating with basis functions
- difference between ML decision and MAP decision
- meaning of simplex signals
- difference between the number of coordinates and the dimension

여기까지 시험범위입니다.

10월 25일(수10,11) QUIZ#2 -- Chapter 3 의 일부분 + 수업내용 + HW 3 &4

THIRD QUARTER (GO TO 1st 2nd 3rd 4th quarter; top today bottom )

11월 1일(수8)

2차시험 문제풀이 - 답안지확인
N-variate joint Gaussian random variables

11월 1일(수10,11) 보강

Non-coherent Detection of Binary Signals in AWGN
- model
- Generalized ML decision
- Optimum non-coherent receiver structure
- Performance Analysis
HW#5 - Performance analysis of Non-coherent Binary Detection over AWGN: due 11월 15일 (수)

11월 3일(금78)

DPSK encoding and decoding
- coherent receiver: roughly twice the BER of coherent BPSK signaling
- noncoherent optimum receiver: same as optimum noncoherent reception of orthogonal signals with energy 2E
- suboptimum noncoherent receiver ( = differentially coherent detection of DPSK): almost the same as coherent reception of orthogonal signals.
Summary : Binary, perfect timing (and phase for coherent reception), equal energy, equally-likely a priori, AWGN with 2-sided PSD=N0/2
- Coherent opt. reception of BPSK: BER = Q(sqrt{2E/N0})
- Coherent opt. reception of optimum BFSK: BER = Q(sqrt{(E(1+2/3pi)/N0)})
- Coherent opt. reception of orthogonal BFSK: BER = Q(sqrt{E/N0})
- Noncoherent opt. reception of orthogonal BFSK: BER = (1/2) exp ( -E/2N0)
- Coherent opt. reception of DPSK: BER = 2Q(sqrt{2E/N0})
- Noncoherent opt. reception of DPSK: BER = (1/2) exp ( -E/N0)
- Differentially coherent reception of DPSK (= suboptimum noncoherent reception of DPSK): BER = Q(sqrt{ (E/N0) / [1+ (1/(4E/N0)) ] })
Self-exercise: Plot BER curves vs E/N0 using the table of Q-function or some computer program
Professor's suggestion: Completely understand
- various binary waveforms,
- optimum receiver's block diagram,
- how to derive BER performance and required assumptions,
- various BER relation between different methods, and
- memorize Fig 3.22 of the text !!
relation of simplex signals with orthogonal signals
decoding with non-white noise
- whitening filter followed by MF matched to whitened signal
- BER does depend on the actual waveshape of binary signals. (previously, it depends only on its energy and corr coeff.)
- how to choose optimum waveforms ?

11월 8일(수8) 조교수업 합니다.

11월 8일(수10) 연습 없음

11월 10일(금78) 조교수업 합니다.

Topic: "Concept of Signal Space" or something else

from Communications Handbook by Jerry Gibbson

11월 15일(수8)

Review Summary : Binary, perfect timing (and phase for coherent reception), equal energy, equally-likely a priori, AWGN with 2-sided PSD=N0/2
- Coherent opt. reception of BPSK: BER = Q(sqrt{2E/N0})
- Coherent opt. reception of optimum BFSK: BER = Q(sqrt{(E(1+2/3pi)/N0)})
- Coherent opt. reception of orthogonal BFSK: BER = Q(sqrt{E/N0})
- Noncoherent opt. reception of orthogonal BFSK: BER = (1/2) exp ( -E/2N0)
- Coherent opt. reception of DPSK: BER = 2Q(sqrt{2E/N0})
- Noncoherent opt. reception of DPSK: BER = (1/2) exp ( -E/N0)
- Differentially coherent reception of DPSK (= suboptimum noncoherent reception of DPSK): BER = Q(sqrt{ (E/N0) / [1+ (1/(4E/N0)) ] })
HW#6 part 1
- For the 7 types of binary signaling above,
  1. Specify the signal set (waveform)
  2. describe the receiver block diagram
  3. write a program to calculate BER given above and plot 7 curves corresponding to each scheme.
- The curves must show the relation between BER and Eb/No for approximately BER=10E-7 or Eb/No upto 12 dB.
- What you must hand in by 29th of Nov:
  1. Description of each scheme (1 and 2 above)
  2. Program code printout, and BER calculation results in table, and graph containing 7 curves.

decoding with non-white noise
KL-expansion of random processes
- A random process has KL expansion if its correlation function satisfies certain integral equation.
- KL expansion is an orthonormal basis expansion of a random process in which
- (1) coefficients are uncorrelated random variables and
- (2) orthonormal basis functions are in fact eigen functions of correlation function of the random process, and
- (3) average energy is the sum of the eigenvalues.

11월 15일(수) 연습시간 없음 - 지도교수 간담회 참석요망

11월 17일(금78)

optimum pair of antipodal signals for non-white noise
Coherent MPSK
orthogonal MFSK
HW#6 part 2 --Chapter 3 연습문제 중에서--- 4,5,6,7,11,12,13,14,15,16 (17,18,19는 제외)
HW#6 Due date = 11월 29일(수) 수업시간시작시.

FOURTH QUARTER (GO TO 1st 2nd 3rd 4th quarter; top today bottom )

11월 22일(수8) Chapter 4: Communication Link Analysis

range equation
thermal noise and noise power spectral density
req(Eb/No) vs rec(Eb/No) and link margin
noise figure and effective temperature
link analysis
HW#7 --Chapter 4 연습문제 중에서--- 1,2,6,8,9,14,17,18,19,20
HW#7 Due date = 12월 6일(수) 수업시간시작시.

학기말까지의 강의노트를 복사실에.
- error correcting block codes
- convolution codes
- spread spectrum communications
- pseudonoise sequences
- introduction to IS95 air-interface

11월 22일(수 10,11) Discussion on HW#4 and HW#5

11월 24일(금78) Error Correction Codes - block codes

source coding, channel coding, and secrecy coding
concept of ECC
example: BSC with repetition code and Message Error Probability
Shannon's Noisy Coding Theorem
random error and burst error
FEC, ARQ, Error Concealment
complete/incomplete decoding
linear/nonlinear, block/tree, binary/nonbinary
Hamming codes, BCH codes, RS codes, NR codes, etc

11월 29일(수8) Error Correction Codes

error-correction capability
minimum distance, minimum weight
[3,1,3] hamming code = 3-time-repetition code

HW#8 part 1 --Chapter 5 연습문제 중에서 - 2,3,4,5,8,11,16,25,26,27
HW#8 part 2 --Chapter 6 연습문제 중에서 - 1,5,6,8,9,10,11,15,17,18 -- 이 부분은 강의노트의 연습문제 I, II, III으로 대치합니다.
HW#8 Due date = 12월 13일(수) 수업시간시작시. (no more HW. this is final)

11월 29일(수 10,11) Discussion on HW#6 Part 2

12월 1일(금78) Error Correction Codes

complete decoding vs incomplete decoding
fundamental problem of (linear) coding theory
relation with n, k, and d

12월 2일(토: 2시-5시, 보강)

binary hamming codes
generator matrix, parity check matrix
cyclic code
systematic encoding
shortening
extended hamming codes
standard array decoding
hard/soft decision decoding
uncoded MPSK vs coded MPSK systems
coding gain

12월 6일(수8) convolution codes

강좌평가설문, 온라인 홈페이지 평가설문
convolution code 개요
encoder block diagram (example) and parameters
- code rate, constraint length, minimum free distance
state diagram
distance structure
viterbi decoding algorithm is ML decoding
LONG EXAMPLE of viterbi decoding

12월 6일(수 10,11) Discussion on HW#7

12월 8일(금78) Spread Spectrum Communications

Historical Background
concept of Processing Gain
DS vs FH
Broadband Noise Jamming vs partialband Jamming
PN sequence(code) acquisition and tracking is important
Application to CDMA Cellular Communication

12월 13일(수8) - 종강 - PN sequences

m-sequence
autocorrelation property
balance property
span-n property
etc

12월 13일(수 10,11) Discussion on HW#8

12월 15일(금) 기말시험 (Quiz #3 and Quiz #4) 오후 8시-10시 B004

12월 22일(금) 성적처리전 공고

Go Back Home

(GO TO 1st 2nd 3rd 4th quarter; top today bottom )