review of fourier series - georgia institute of...

21
1 Copyright Gregory D. Durgin 2000 VIRGINIA POLYTECHNIC INSTITUTE AND STATE UNIVERSITY Tech V irginia 1 8 7 2 ECPE 3614 Introduction to Communication Systems Review of Fourier Series ECPE 3614 Lecture 2 Instructor: Gregory D. Durgin Copyright Gregory D. Durgin 2000 VIRGINIA POLYTECHNIC INSTITUTE AND STATE UNIVERSITY Tech V irginia 1 8 7 2 ECPE 3614 Introduction to Communications Systems L2 2 Outline of Lecture § Review of Lecture 1 § Calculation of Fourier Coefficients § Trigonometric vs. Complex FS § Example 1: Triangular Waveform § Example 2: Sawtooth Waveform § Example 3: Rectangular Waveform

Upload: vuonghuong

Post on 26-Apr-2018

217 views

Category:

Documents


2 download

TRANSCRIPT

1

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunication Systems

Review of Fourier Series

ECPE 3614Lecture 2

Instructor: Gregory D. Durgin

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L22

Outline of Lecture

§Review of Lecture 1§Calculation of Fourier Coefficients§Trigonometric vs. Complex FS§Example 1: Triangular Waveform§Example 2: Sawtooth Waveform§Example 3: Rectangular Waveform

2

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L23

Review

§Fourier’s Theorem: All real periodic functionsmay be expressed as a linear combination ofharmonic sinusoids.§There are two types of Fourier Series:§ Trigonometric Fourier Series – uses a series sines and

cosines.§ Complex Fourier Series – uses a series of complex

exponentials as oscillators.§ These series are just two different ways of doing the

same thing.

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L24

The Trigonometric FourierSeries

§Mathematically, the trigonometric Fourier seriesmay be written in this form:

§Note that the n=0 case for the sine summation isnot important since the term drops out to zero

3

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L25

Calculating TrigonometricFourier Series Coefficients

Let’s say we want to calculate the mth cosine coefficient, am.How do we go about it? First, start with the definition of thetrig. FS:

Step 1: multiply both sides by the mth harmonic cosine asshown below:

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L26

Calculating TrigonometricFourier Coefficients

Step 2: Integrate both sides over exactly one period of theperiodic function. (This is starting to look complicated, butwe can simplify in a few steps.)

4

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L27

Calculating TrigonometricFourier Coefficients

Step 3: The right-side integral may be distributed across eachterm in the series summation:

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L28

Calculating TrigonometricFourier Coefficients

Step 4: Take a look and the cos-sin integral:

This is a standard definite integral that can be found in manybooks. Since this is an integration of a harmonic sine wave (nis an integer) and a harmonic cosine wave (m is an integer)over exactly one period, T, the integral always evaluates to 0.

5

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L29

Calculating TrigonometricFourier Coefficients

Step 5: There are 3 possible outcomes for the cos-cos definiteintegral:

§m and n are different: the integral always evaluates to 0.§m and n are the same (but not 0): the integral evaluates to

T/2.§m and n are 0: the integral evaluates to T.

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L210

Calculating TrigonometricFourier Coefficients

Step 6: Good news! Based on the definite integrations ofsteps 4 and 5, all but one of the terms will vanish.

always 0

always 0 except forthe case m=n.

6

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L211

Calculating TrigonometricFourier Coefficients

Step 7: For the case of m>0, the results of our hard effortsproduces an equation for am:

Just switch the m to an n (to keep notation consistent) andrearrange the equation to solve for an. Now we have theformula for the nth cosine Fourier coefficient!

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L212

Calculating TrigonometricFourier Coefficients

Step 8: One last step – for the case of m=0, the results of ourhard efforts produces an equation for a0:

Rearranging, the equation for a0 becomes…

7

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L213

Calculating the TrigonometricFourier Coefficients…

The full set of trigonometricFourier coefficients is shown on theright. The derivation for the bn isidentical to the an, except that asine term is used in Step 1. Note:the limits of integration for theseformulas are somewhat arbitrary.Since the integrands are periodicfunctions, any integration lengththat spans exactly one period, T,will produce the correct result.

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L214

Example of Trig FS:A Triangular Waveform

§Blackboard Example: Find the trigonometric Fouriercoefficients for this triangular waveform with period of 1s.

time, t

+1/2

1 s

8

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L215

The Complex Fourier Series

§Below is the complex Fourier Series:

§Key points about the complex series:§ unlike an and bn, cn may be a complex number.§ summation limits are from –infinity to +infinity.§ if f(t) is to be a real-valued function, the complex

Fourier coefficients must have special properties (seeHomework 1, problem 1).

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L216

Calculating The ComplexFourier Coefficients

Start with the definition of the complex Fourier Coefficients:

Step 1: To calculate the mth coefficient, multiply both sidesof the equation by the –mth complex exponential harmonic:

9

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L217

Calculating The ComplexFourier Coefficients

Step 2: Integrate both sides over an entire period, T.

Note: The imaginary j does not change how you performintegration. Just take it along for the ride, as if it were aconstant.

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L218

Calculating The ComplexFourier Coefficients

Step 3: Distribute the integral to each term inside thesummation. This step is valid because integration is linear.

10

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L219

Calculating The ComplexFourier Coefficients

Step 4: Study the following integral:

Since this is an integration of complex harmonic sinusoids(the value n-m is an integer) over exactly one period, theresult is only non-zero for m=n. For m=n, the integralevaluates to T.

0 unless m=n

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L220

Calculating The ComplexFourier Coefficients

Step 5: After all of the zero terms are removed, the finalresult is given below:

Swapping the letter m for n (for consistency of notation) andre-arranging provides us the formula for the complex Fouriercoefficients:

11

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L221

The Complex Fourier Series

Below is the integral for calculating complex Fourier coefficients.The drawback to this method is the integration of the complexexponent. This integration often produces a complex coefficientfor cn. One nice advantage: only one formula as opposed to the 3equations required for the trigonometric series.

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L222

To Convert from Complex FSto Trig FS Coefficients…

… just make the followingsubstitutions. This cansave a lot of computation!

Remember: the complex Fourier series is justa different way to keep track of the spectral

decomposition.

12

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L223

To Convert from Trig FS toComplex FS Coefficients…

… just make the followingsubstitutions.

The set of harmonic sinusoids (both complexexponential and trigonometric) are said to be abasis that spans the set of real periodic functions.

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L224

Ex. 1: Triangular Waveform

time, t

+1/2

1 s

Trig FS coefficients:

13

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L225

-2 -1.5 -1 -0.5 0 0.5 1 1.5 2-0.5

0

0.5

1Triangle Waveform of T=1

time,t

Sig

nal A

mpl

itude

Exact SignalFourier Series through term 0

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L226

-2 -1.5 -1 -0.5 0 0.5 1 1.5 2-0.5

0

0.5

1Triangle Waveform of T=1

time,t

Sig

nal A

mpl

itude

Exact SignalFourier Series through term 1

14

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L227

-2 -1.5 -1 -0.5 0 0.5 1 1.5 2-0.5

0

0.5

1Triangle Waveform of T=1

time,t

Sig

nal A

mpl

itude

Exact SignalFourier Series through term 3

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L228

-2 -1.5 -1 -0.5 0 0.5 1 1.5 2-0.5

0

0.5

1Triangle Waveform of T=1

time,t

Sig

nal A

mpl

itude

Exact SignalFourier Series through term 5

15

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L229

Ex. 2: The SawtoothWaveform

Trig FS coefficients:

time, t

+1/2

-1/2

1 s

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L230

-2 -1.5 -1 -0.5 0 0.5 1 1.5 2-0.5

0

0.5

1Sawtooth Waveform of T=1

time, t

Sig

nal A

mpl

itude

Exact SignalFourier Series through term 1

16

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L231

-2 -1.5 -1 -0.5 0 0.5 1 1.5 2-0.5

0

0.5

1Sawtooth Waveform of T=1

time, t

Sig

nal A

mpl

itude

Exact SignalFourier Series through term 3

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L232

-2 -1.5 -1 -0.5 0 0.5 1 1.5 2-0.5

0

0.5

1Sawtooth Waveform of T=1

time, t

Sig

nal A

mpl

itude

Exact SignalFourier Series through term 6

17

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L233

-2 -1.5 -1 -0.5 0 0.5 1 1.5 2-0.5

0

0.5

1Sawtooth Waveform of T=1

time, t

Sig

nal A

mpl

itude

Exact SignalFourier Series through term 15

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L234

Ex. 3: The RectangularWaveform

Trig FS coefficients:

time, t

+1/2

1 s

18

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L235

-2 -1.5 -1 -0.5 0 0.5 1 1.5 2-0.5

0

0.5

1Rectangular Waveform of T=1

time,t

Sig

nal A

mpl

itude

Exact SignalFourier Series through term 1

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L236

-2 -1.5 -1 -0.5 0 0.5 1 1.5 2-0.5

0

0.5

1Rectangular Waveform of T=1

time,t

Sig

nal A

mpl

itude

Exact SignalFourier Series through term 3

19

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L237

-2 -1.5 -1 -0.5 0 0.5 1 1.5 2-0.5

0

0.5

1Rectangular Waveform of T=1

time,t

Sig

nal A

mpl

itude

Exact SignalFourier Series through term 5

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L238

-2 -1.5 -1 -0.5 0 0.5 1 1.5 2-0.5

0

0.5

1Rectangular Waveform of T=1

time,t

Sig

nal A

mpl

itude

Exact SignalFourier Series through term 7

20

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L239

-2 -1.5 -1 -0.5 0 0.5 1 1.5 2-0.5

0

0.5

1Rectangular Waveform of T=1

time,t

Sig

nal A

mpl

itude

Exact SignalFourier Series through term 13

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L240

-2 -1.5 -1 -0.5 0 0.5 1 1.5 2-0.5

0

0.5

1Rectangular Waveform of T=1

time,t

Sig

nal A

mpl

itude

Exact SignalFourier Series through term 25

21

Copyright Gregory D. Durgin 2000VIRGINIA POLYTECHNIC INSTITUTEAND STATE UNIVERSITY

TechVirginia

1 8 7 2

ECPE 3614 Introduction toCommunications Systems

L241

Key Points of Lecture

§Basic Form of the Fourier Series§Calculate Trigonometric Fourier Coefficients§Calculate Complex Fourier Coefficients§Convert Between Complex and Trig Fourier

Coefficients§Examples of how truncated Fourier Series

approximate real signals with increasing accuracy