M-ary Detection Problem. Lecture Notes 2: Detection Theory. Example 1: Additve White Gaussian Noise

Similar documents
The Non-Truncated Bulk Arrival Queue M x /M/1 with Reneging, Balking, State-Dependent and an Additional Server for Longer Queues

Lecture 25 Outline: LTI Systems: Causality, Stability, Feedback

Physics 232 Exam I Feb. 13, 2006

BINOMIAL THEOREM OBJECTIVE PROBLEMS in the expansion of ( 3 +kx ) are equal. Then k =

Lecture 4. Electrons and Holes in Semiconductors

Cameras and World Geometry

Physics 232 Exam I Feb. 14, 2005

Two-Pion Exchange Currents in Photodisintegration of the Deuteron

Optical flow equation

Lecture 4. Electrons and Holes in Semiconductors

Angle Modulation: NB (Sinusoid)

Comparing Different Estimators for Parameters of Kumaraswamy Distribution

Lesson 5. Chapter 7. Wiener Filters. Bengt Mandersson. r x k We assume uncorrelated noise v(n). LTH. September 2010

Strong Result for Level Crossings of Random Polynomials. Dipty Rani Dhal, Dr. P. K. Mishra. Department of Mathematics, CET, BPUT, BBSR, ODISHA, INDIA

CODING & MODULATION Prof. Ing. Anton Čižmár, PhD. also from Digital Communications 4th Ed., J. G. Proakis, McGraw-Hill Int. Ed.

ABSOLUTE INDEXED SUMMABILITY FACTOR OF AN INFINITE SERIES USING QUASI-F-POWER INCREASING SEQUENCES

ECSE Partial fraction expansion (m<n) 3 types of poles Simple Real poles Real Equal poles

Meromorphic Functions Sharing Three Values *

Exercise: Show that. Remarks: (i) Fc(l) is not continuous at l=c. (ii) In general, we have. yn ¾¾. Solution:

Consider a Binary antipodal system which produces data of δ (t)

Strong Result for Level Crossings of Random Polynomials

Physics 232 Exam II Mar. 28, 2005

Outline. Review Homework Problem. Review Homework Problem II. Review Dimensionless Problem. Review Convection Problem

Introduction to Hypothesis Testing

New Results on Oscillation of even Order Neutral Differential Equations with Deviating Arguments

Available online at J. Math. Comput. Sci. 2 (2012), No. 4, ISSN:

The Nehari Manifold for a Class of Elliptic Equations of P-laplacian Type. S. Khademloo and H. Mohammadnia. afrouzi

X-Ray Notes, Part III

State-Space Model. In general, the dynamic equations of a lumped-parameter continuous system may be represented by

AB for hydrogen in steel is What is the molar flux of the hydrogen through the steel? Δx Wall. s kmole

, on the power of the transmitter P t fed to it, and on the distance R between the antenna and the observation point as. r r t

Then the number of elements of S of weight n is exactly the number of compositions of n into k parts.

Mathematical Models and the Soil Hydraulic Properties

Comments on Discussion Sheet 18 and Worksheet 18 ( ) An Introduction to Hypothesis Testing

ME 321 Kinematics and Dynamics of Machines S. Lambert Winter 2002

t = s D Overview of Tests Two-Sample t-test: Independent Samples Independent Samples t-test Difference between Means in a Two-sample Experiment

Supplementary Information

CHATTERJEA CONTRACTION MAPPING THEOREM IN CONE HEPTAGONAL METRIC SPACE

Review - Week 10. There are two types of errors one can make when performing significance tests:

Summary of Grade 1 and 2 Braille

CHAPTER 2 Quadratic diophantine equations with two unknowns

Maximum Likelihood Estimation

Chapter 2: Descriptive Statistics

S.E. Sem. III [EXTC] Applied Mathematics - III

ECE 350 Matlab-Based Project #3

. Since P-U I= P+ (p-l)} Aap Since pn for every GF(pn) we have A pn A Therefore. As As. A,Ap. (Zp,+,.) ON FUNDAMENTAL SETS OVER A FINITE FIELD

Last time: Completed solution to the optimum linear filter in real-time operation

Revision of Lecture Eight

ON THE EXTENSION OF WEAK ARMENDARIZ RINGS RELATIVE TO A MONOID

Clicks, concurrency and Khoisan

Communications II Lecture 4: Effects of Noise on AM. Professor Kin K. Leung EEE and Computing Departments Imperial College London Copyright reserved

FI 2201 Electromagnetism

Congruences for sequences similar to Euler numbers

Lesson 5. Chapter 7. Wiener Filters. Bengt Mandersson. r k s r x LTH. September Prediction Error Filter PEF (second order) from chapter 4

EQUATION SHEET Principles of Finance Exam 1

rad / sec min rev 60sec. 2* rad / sec s

TRAVELING WAVES. Chapter Simple Wave Motion. Waves in which the disturbance is parallel to the direction of propagation are called the

10.3 Autocorrelation Function of Ergodic RP 10.4 Power Spectral Density of Ergodic RP 10.5 Normal RP (Gaussian RP)

Chapter 9. Key Ideas Hypothesis Test (Two Populations)

Moment Generating Function

Neutron Slowing Down Distances and Times in Hydrogenous Materials. Erin Boyd May 10, 2005

Secure Chaotic Spread Spectrum Systems

Lecture 15: Three-tank Mixing and Lead Poisoning

Construction of Malliavin differentiable strong solutions of SDEs under an integrability condition on the drift without the Yamada-Watanabe principle

10-716: Advanced Machine Learning Spring Lecture 13: March 5

Low-complexity Algorithms for MIMO Multiplexing Systems

th m m m m central moment : E[( X X) ] ( X X) ( x X) f ( x)

LIPSCHITZ ESTIMATES FOR MULTILINEAR COMMUTATOR OF MARCINKIEWICZ OPERATOR

Hadamard matrices from the Multiplication Table of the Finite Fields

Hidden Markov Model Parameters

8.6 Order-Recursive LS s[n]

Lecture 5. Chapter 3. Electromagnetic Theory, Photons, and Light

Valley Forge Middle School Fencing Project Facilities Committee Meeting February 2016

EEC 483 Computer Organization

Topics in MMSE Estimation for Sparse Approximation

Construction of Malliavin differentiable strong solutions of SDEs under an integrability condition on the drift without the Yamada-Watanabe principle

Chapter 15: Fourier Series

Two Implicit Runge-Kutta Methods for Stochastic Differential Equation

S n. = n. Sum of first n terms of an A. P is

Pattern Distributions of Legendre Sequences

هقارنت طرائق تقذير هعلواث توزيع كاها ري املعلوتني

Dividing Algebraic Fractions

Processamento Digital de Sinal

BEST LINEAR FORECASTS VS. BEST POSSIBLE FORECASTS

Degree of Approximation of Fourier Series

Spectrum of The Direct Sum of Operators. 1. Introduction

Addition & Subtraction of Polynomials

Test 2 phy a) How is the velocity of a particle defined? b) What is an inertial reference frame? c) Describe friction.

LECTURE 13 SIMULTANEOUS EQUATIONS

An interesting result about subset sums. Nitu Kitchloo. Lior Pachter. November 27, Abstract

Stat 3411 Spring 2011 Assignment 6 Answers

TIME RESPONSE Introduction

Physics 201 Lecture 15

EECE 301 Signals & Systems Prof. Mark Fowler

u t u 0 ( 7) Intuitively, the maximum principles can be explained by the following observation. Recall

The shortest path between two truths in the real domain passes through the complex domain. J. Hadamard

ASTR 3740 Relativity & Cosmology Spring Answers to Problem Set 4.

Math 213b (Spring 2005) Yum-Tong Siu 1. Explicit Formula for Logarithmic Derivative of Riemann Zeta Function

P a g e 3 6 of R e p o r t P B 4 / 0 9

SLOW INCREASING FUNCTIONS AND THEIR APPLICATIONS TO SOME PROBLEMS IN NUMBER THEORY

Transcription:

Hi ue Hi ue -ay Deecio Pole Coide he ole of decidig which of hyohei i ue aed o oevig a ado vaiale (veco). he efoace cieia we coide i he aveage eo oailiy. ha i he oailiy of decidig ayhig ece hyohei H whe hyohei H i ue. Lecue oe : Deecio heoy he udelyig odel i ha hee i a codiioal oailiy deiy (a) fucio of he oevaio give each hyohei H. Goal: Oiu Deecio i AWG Oiu Deecio wih uiace (Uwaed) Paaee Pe P i decide H P decide H i i Ri i i Ri i πid πid II- II- he deciio ule ha iiize aveage co aig o R i if i a πi Le e a aiay deiy fucio ha i ozeo eveywhee i i ozeo he a equivale deciio ule i o aig o R i if i a hu fo hyohee he deciio ule ha iiize aveage eo oailiy i o chooe i o ha i i. Le whee πi π, Λ i Chooe i if Λ i i π. he he oial deciio ule i: π fo all i We will uually aue i. (If o we hould do ouce ecodig o educe he eoy (ae)). Fo hi cae he oial deciio ule i Chooe i if Λ i i π. ale : Addive Whie Gauia oie Coide hee igal i addiive whie Gauia oie. Fo addiive whie Gauia oie K δ. Leϕi e ay colee ohooal e o. Coide he cae of 3 igal. Fid he deciio ule o iiize aveage eo oailiy. Fi ad he oie uig ohooal e of fucio ad ado vaiale. i ϕ i whee i ad Va i ad i i a ideede ideically diiued (i.i.d.) equece of ado vaiale wih Gauia deiy fucio. Le ϕ ϕ oe ha he eegy of each of he hee igal i he ae, i.e. ϕ ϕ ϕ ϕ i d i 5. he II-3 II-4

we have a hee hyohei eig ole. H : H : H : i i i i i i he deciio ule o iiize he aveage eo oailiy i give a follow Decide H i if i a π Fie u oalize each ide y he deiy fucio fo he oie aloe. he oie deiy fucio fo vaiale i π he he oial deciio ule i equivale o i Decide H i if a π i ϕi ϕi ϕi A uual aue. he π π i 4 5 i i i i i i i ow ice he aove doe deed o we ca le adhe eul i he ae, i.e. Siilaly li i 4 5 4 5 4 5 II-5 II-6 φ Deciio Regio 8 6 4 φ 4 6 8 8 6 4 4 6 8 II-7 II-8

ale : Oiu Deecio of -ay ohogoal igal fo iiu i eo oailiy I hi ecio we coide he ole of deecio wih uwaed aaee. o illuae coide he ole of iiizig he i eo oailiy i a -ay ohogoal igal e. Le e ohogoal igal. he eceive coi of a a of ached file (coelao) ha geeae a ufficie aiic. If igal i aied he δ δ η η φ φ δ η φ Le e he equece of i deeiig which of he igal i aied. Aue he i ae ideede ad equally liely. Coide he deecio of daa i. ha i, we ae ieeed i iiizig he oailiy of eo fo daa i. Le H e he eve ha ad H e he eve ha. Le. he he oial eceive u coae he wo aoeioi oailiie H π H H H π II-9 II- o calculae H we oceed a follow. H Siilaly π π πσ πσ πσ πσ π πσ πσ σ σ σ σ σ σ l l l l δl δl l l lδl δl σ σ σ l δ l σ H π πσ π oice ha ay of he faco i aio fo i he log-lielihood aio i log H H π π hi ca e aoiaed y log H H π π H H π π H log a σ l π ad σ σ H σ σ π ae he ae. hu he lielihood σ σ log a σ σ II- II-

ale 3: Oiu Deecio of iay igal i fadig chael Coide a ye wih L aea. Aue ha he eceive ow eacly he faded aliude o each aea. he deciio aiic ae he give y l z l η l l whee l ae Rayleigh, η l i Gauia ad eee he daa i aied which i eihe + o -. he ado vaiale l eee he fadig fo he aie o he l deiy l σ e σ L h aea ad ha We aue he fadig o each aea i ideede. he oial ehod o coie he deodulao ouu ca e deived a follow. Le z z L L e he codiioal deiy fucio of z z L give he aied i i + ad he fadig aliude i L. he ucodiioal deiy i z z L L z z L he codiioal deiy of z give ad, i Gauia wih ea l. he oi diiuio of z z L i he oduc of he agial deiy fucio. he oial coiig ule i deived fo he aio Λ z z z z z z z L z L z L z L z L z L L l z l 4 L l L l L L L L L L L l zl l z l l L L L ad vaiace II-3 II-4 he oiu deciio ule i o coae Λ wih o ae a deciio. hu he oial ule i L l l z l oe ha we do o eed o ow he deiy of he aliude fo hi deciio ule. hi deciio ule i called aiu aio coiig (RC). I he ecial cae whee hee i u oe aea he oiu eceive educe o z hu he oiu eceive fo u oe aea (ad BPSK) doe o eed he ifoaio aou he eceived aliude o ae a (had) deciio. Howeve, he efoace deed ciically o he diiuio of he fadig aliude. Fo he Rayleigh faded cae he eo oailiy ecoe P e Ē Ē z Lielihood Raio fo Real Sigal i AG Aue wo igal i Gauia oie. H : H : Goal: Fid deciio ule o iiize he aveage eo oailiy. Le have covaiace K eigefucio ϕi wih ad eigevalue. We aue ha i a zeo ea Gauia ado oce. he eige fucio ϕ i ae ohooal fucio ad eal ue uch ha (ee Aedi) K d ϕi λiϕi II-5 II-6

d d By Kahue-Loeve aio Kahue-Loeve aio i i ϕ i whee i i Gauia ea i vaiace. i i i π πλi i i i i whee i ae Gauia ado vaiale wih ea vaiace ad i i ideede i eal). Sice ϕ i ae a colee ohooal e ad we aue ha fiie eegy we have hu Defie H : Λ i iϕ i. Λ i i i i i i i li ϕi i Λ i Le Λ l l li πλi πλi i i i ϕ i i i i l i i i i iϕ i i i l i i i i i l i II-7 II-8 he hu Λ l li Λ l i i l ϕ l i l i i ϕ i d l iϕ i ϕ l d ϕ i ql ϕl d ql So K d i ϕ i K K d If he oie i whie, he he oie owe i each diecio i coa (ay λ) ad hu he oial eceive he ecoe o equivalely Λ l λ i λ ϕ i λ oe: i oluio of he iegal equaio Λ l λ II-9 II-

Lielihood Raio fo Cole Sigal Fo equal eegy igal hi aou o icig he igal wih he lage coelaio wih he eceived igal. he oial eceive i owhie Gauia oie ca e ileeed i a iila fahio a how elow. hu Λ l K K K K K K K K K K K I i clea he ha hi iu he oial file fo igal K whe eceived i addiive whie Gauia oie. hi aoach i called whieig ecaue K will e a whie Gauia oie oce. I hi ecio we edeive he lielihood aio fo cole igal eceived i cole oie. We aue ha he igal ae he lowa eeeaio of ada igal ad he oie i he lowa eeeaio of a aowad ado oce. Le H : H : whee ha covaiace K, wih eigefucio ϕ i, eigevalue. Uig Kahue-Loeve aio we have i πλ e l e π H i : l l l i l i l l λl λl i l il ϕ λl II- II- Le i ϕ i he l l l l l i l l Re Re i l l a l l ϕ l λl l l l i l λl a l l l ϕ l λ l ϕ ϕ d d i l Re l i l So Λ i li H Λ i i i oe: Sice we ae dealig wih oie ha i deived fo a aowad ado oce we ca o ue he eul deived fo eal ado ocee we u ue he lielihood aio fo cole ado oce give aove. Fo eal ado oce he lielihood aio i Λ i Fo addiive whie Gauia oie (eal) q i i ϕ λ i Re i i ϕ q II-3 II-4

So he lielihood aio (fo eal igal) ecoe Λ lli l H H l ale: ohogoal igal i addiive whie Gauia oie I hi ecio we coide he oiu eceive fo -ay ohogoal igal ad he aociaed eo oailiy. Aue he igal ae equieegy igal ad equioale. he deciio ule deived eviouly fo AWG i Decide H i if i i Aue π. he α. A equivale deciio ule he i H l H H l H he oiu deciio ule fo addiive whie Gauia oie i he o chooe i if i i ow ice he igal ae ohogoal ad equieegy we ca wie hi a he fi e aove i coa fo each a i he la e. hu fidig he iiu i equivale o fidig he aiu of hu he eceive hould coue he ie oduc ewee he diffee igal ad fid he lage uch coelaio. If he igal ae all of duaio, i.e. zeo ouide he ieval he hi i alo equivale o fileig he eceived igal wih a file wih ilue eoe, alig he ouu of he file a ie ad chooig he lage a how elow. II-5 II-6