Linear Controller Design: Limits of Performance by Stephen Boyd and Craig Barratt - HTML preview

PLEASE NOTE: This is an HTML preview only and some elements such as links or page numbers may be incorrect.
Download the book in PDF, ePub, Kindle for a complete version.

CHAPTER 4 NORMS OF SIGNALS

1

0:8

) 0:6

(a

u

F

0:4

0:2

u aa

k

k

rms

k

u

k

k

u

1

k

A

A

U

?

0

0

0:5

1

1:5

2

2:5

3

a

The amplitude distribution

of the signal u shown in g-

Figure

4.10

F

u

ure 4.1, together with the values u aa, u rms, and u .

k

k

k

k

k

k

1

From elementary probability theory we have

Z

1

aa =

( ) =

( )

(4.9)

kuk

E

ju

t

j

0 F a da:

u

Thus, the average-absolute norm of a signal is the total area under the amplitude

distribution function.

Since the amplitude distribution function of 2 is

( ) = (p ), equa-

2

u

F

a

F

a

u

u

tion (4.9) yields

Z

Z

Z

( )2 = 1

( ) = 1 (

1

p

) =

2 ( )

2

E

u

t

0 F a da

a

da

aF

a

da

u

u

0 Fu

0

so that we can express the RMS norm as:

2

Z

1

rms =

2 ( )

(4.10)

kuk

0

aF

a

da:

u

Thus, the average power in the signal is the integral of its amplitude distribution

function times 2 . Just as we interpret formula (4.6) as expressing the average

a

power in the signal as the integral of the contributions at all frequencies, we may

interpret (4.10) as expressing the average power in the signal as the integral of the

contributions from all possible signal amplitudes.

Comparison of the three formulas (4.8), (4.9), and (4.10) show that the three

norms simply put di erent emphasis on large and small signal values: the steady-

state peak norm puts all of its emphasis on large values the RMS norm puts

index-88_1.png

index-88_2.png

index-88_3.png

index-88_4.png

4.2 COMMON NORMS OF SCALAR SIGNALS

79

1

2

0 8:

1

0 6

(t)

(a) :

u1 0

1

u

0 4:

1

F

;

0 2:

2

;

0

2

4

0

t 6

8

10

0

0 4

0 8

1 2

1 6

2

:

:

a :

:

(a)

(b)

1

4

0 8:

0 6

(t) 2

(a) :

u2

2

u

0 4:

F

0

0 2:

2

0

;

0

2

4 t 6

8

10

0

1

a2

3

4

(c)

(d)

Examples of periodic signals are shown in (a) and (c). Their

Figure

4.11

respective amplitude distribution functions are shown in (b) and (d). The

signal in (a) spends most of its time near its peaks the amplitude distribu-

tion falls rapidly near a = u1 . The signal in (c) spends most of its time

k

k

1

near 0 the amplitude distribution falls rapidly near a = 0.

index-89_1.png

index-89_2.png

index-89_3.png

index-89_4.png

80