6.

003: Signals and Systems

Discrete Approximation of Continuous-Time Systems

September 29, 2011

1

Mid-term Examination #1
Wednesday, October 5, 7:30-9:30pm, No recitations on the day of the exam. Coverage: CT and DT Systems, Z and Laplace Transforms Lectures 17 Recitations 17 Homeworks 14

Homework 4 will not collected or graded. Solutions will be posted. Closed book: 1 page of notes (8 1 2 11 inches; front and back). No calculators, computers, cell phones, music players, or other aids. Designed as 1-hour exam; two hours to complete. Review sessions during open oce hours. Conict? Contact before Friday, Sept. 30, 5pm. Prior term midterm exams have been posted on the 6.003 website.
2

Concept Map
Today we will look at relations between CT and DT representations.
Delay R

Block Diagram
X +
Delay

System Functional
Y

+
Delay

Y 1 = H(R) = X 1 R R2

DT
index shift

Unit-Sample Response

h[n] : 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, . . .

Dierence Equation

CT
Block Diagram
X +

y [n] = x[n] + y [n 1] + y [n 2]

H (z ) =

System Function Y (z ) z2 = 2 X (z ) z z1

Delay R

System Functional
Y

+ 1

1 2

Y 2A2 = X 2 + 3A + A2

T D

Impulse Response
x (t) x(t)

h(t) = 2(et/2 et ) u(t)

CT
Dierential Equation

2 y (t) + 3y (t) + y (t) = 2x(t)

System Function Y (s) 2 = 2 X (s) 2s + 3s + 1

Discrete Approximation of CT Systems

Example: leaky tank
r0 (t)

h1 (t) r1 (t)
X AX

Block Diagram
X +
1

System Functional
Y

A Y = X A+

Impulse Response
x (t) x(t)
1 et/ u(t) h(t) =

Dierential Equation

r 1 (t) = r0 (t) r1 (t)

H (s) =

System Function Y (s) 1 = X (s) 1 + s

Today: compare step responses of leaky tank and DT approximation.

4

Check Yourself (Practice for Exam)

What is the step response of the leaky tank system?

u(t)

Leaky Tank

s(t) =?

1
1. 2.

1 t 1
4.

1
3.

5. none of the above

5

Check Yourself
What is the step response of the leaky tank system? de: r 1 (t) = u(t) r1 (t)

t < 0: r1 (t) = 0 t > 0: r1 (t) = c1 + c2 et/ r 1 (t) = c2 et/ Substitute into de: c2 et/ = 1 c1 c2 et/ Combine t < 0 and t > 0: r1 (t) = u(t) + c2 et/ u(t) c2 r 1 (t) = (t) + c2 (t) et/ u(t) Substitute into de: c2 (1 + c2 ) (t) et/ u(t) = u(t) u(t) c2 et/ u(t) r1 (t) = (1 et/ )u(t)
6

c1 = 1

c2 = 1

Check Yourself
Alternatively, reason with systems!

(t) u(t) (t) (t) u(t) A

A A+

A A+ A A+ A A+

1 et/ u(t) h(t) =

s(t) =? s(t) =?
t

h(t) A s(t) =

h(t )dt

s(t) =

t 1 I 1 tI / ' ' u(t )dt = e et / dt' = (1 et/ ) u(t) 0 t

Check Yourself

What is the step response of the leaky tank system?

u(t)

Leaky Tank

s(t) =?

1
1. 2.

1 t 1
4.

1
3.

5. none of the above

8

Forward Euler Approximation

Approximate leaky-tank system using forward Euler approach. Approximate continuous signals by discrete signals: xd [n] = xc (nT ) yd [n] = yc (nT ) Approximate derivative at t = nT by looking forward in time:
y [n +1] yd [n] c (nT ) = d y T

yc (t) yd [n] nT yd [n +1] t

(n +1) T
9

Forward Euler Approximation

Approximate leaky-tank system using forward Euler approach. Substitute xd [n] = xc (nT ) yd [n] = yc (nT ) yc (n + 1)T yc nT y [n + 1] yd [n] = d T T into the dierential equation y c (nT ) y c (t) = xc (t) yc (t) to obtain y [n + 1] yd [n] = xd [n] yd [n] . T d Solve: T T yd [n + 1] 1 yd [n] = xd [n]
10

Forward Euler Approximation

Plot.

1
T

= 0.1

1
T

= 0.3

1
T

=1

1
T

= 1.5

1
T

=2

Why is this approximation badly behaved for large T ? 11

Check Yourself
DT approximation: yd [n + 1] 1 T yd [n] = T x [n] d

Find the DT pole. T 3. z = T 5. z = 1 1+ T T

1. z =

2. z = 1 4. z = T

12

Check Yourself
DT approximation: yd [n + 1] 1 T yd [n] = T x [n] d

Take the Z transform: T T Yd (z ) = Xd (z ) zYd (z ) 1 Solve for the system function: T Y (z ) H (z ) = d = Xd (z ) z 1 T

Pole at z = 1

T .

13

Check Yourself
DT approximation: yd [n + 1] 1 T yd [n] = 2 T T x [n] d

Find the DT pole. T 3. z = T

1. z =

2. z = 1 4. z = 5. z = 1 1+ T T

14

Dependence of DT pole on Stepsize

z 1 t 1 t 1 t 1 t 1 t
T T T T T

= 0.1 z = 0.3 z =1 z = 1.5 z =2

1 ). Why is the DT pole changing? The CT pole was xed (s =

15

Dependence of DT pole on Stepsize

Dependence of DT pole on T is generic property of forward Euler. Approach: make a systems model of forward Euler method. CT block diagrams: adders, gains, and integrators:

X
y (t) = x(t) Forward Euler approximation: y [n + 1] y [n] = x[n] T Equivalent system:

Forward Euler: substitute equivalent system for all integrators.

16

Example: leaky tank system

Started with leaky tank system:

Replace integrator with forward Euler rule:

Write system functional: T R TR TR Y R = 1 = = T T R X 1R+ R 1 + 1R 1 1 T R Equivalent to system we previously developed: T T yd [n] = xd [n] yd [n + 1] 1
17

Model of Forward Euler Method

Replace every integrator in the CT system

X
with the forward Euler model:

Substitute the DT operator for A: T 1 TR T A= = z 1 = z1 s 1R 1 z Forward Euler maps s z1 . T Or equivalently: z = 1 + sT .

18

Dependence of DT pole on Stepsize

Pole at z = 1 T = 1 + sT .

z
T

1 t 1 t 1 t 1 t 1 t
19

= 0.1 z = 0.3 z =1 z

= 1.5 z

=2

Forward Euler: Mapping CT poles to DT poles

Forward Euler Map: s 0
1 T 2 T

z = 1 + sT 1 0 1

1 T 2 T 1 T

s z 1 + sT 1 1

1 at s = 1 . DT stability: CT pole must be inside circle of radius T T

2 1 < <0 T

T <2
20

Backward Euler Approximation

We can do a similar analysis of the backward Euler method. Approximate continuous signals by discrete signals: xd [n] = xc (nT ) yd [n] = yc (nT ) Approximate derivative at t = nT by looking backward in time:
y [n] yd [n 1] y c (nT ) = d T

yc (t) yd [n 1] (n 1)T nT yd [n] t

21

Backward Euler Approximation

We can do a similar analysis of the backward Euler method. Substitute xd [n] = xc (nT ) yd [n] = yc (nT ) yc nT yc (n 1)T y [n] yd [n 1] = d T T into the dierential equation y c (nT ) y c (t) = xc (t) yc (t) to obtain y [n] yd [n 1] = xd [n] yd [n] . T d Solve: T T 1+ yd [n] yd [n 1] = xd [n]

22

Backward Euler Approximation

Plot.

1
T

= 0.1

1
T

= 0.3

1
T

=1

1
T

= 1.5

1
T

=2

This approximation is better behaved. Why?

23

Check Yourself
DT approximation: 1+ T yd [n] yd [n 1] = T x [n] d

Find the DT pole. T 3. z = T 5. z = 1 1+ T T

1. z =

2. z = 1 4. z = T

24

Check Yourself
DT approximation: 1+ T yd [n] yd [n 1] = T x [n] d

Take the Z transform: T T 1+ Yd (z ) z 1 Yd (z ) = Xd (z ) Find the system function: Tz Y (z ) H (z ) = d = T Xd (z ) 1+ z1

Pole at z =

. 1+ T

25

Check Yourself
DT approximation: yd [n + 1] 1 T yd [n] = 5 T T x [n] d

Find the DT pole. T 3. z = T

1. z =

2. z = 1 4. z = 5. z = 1 1+ T T

26

Dependence of DT pole on Stepsize

z 1 t 1 t 1 t 1 t 1 t
T T T T T

= 0.1 z = 0.3 z =1 z = 1.5 z =2

Why is this approximation better behaved?

27

Dependence of DT pole on Stepsize

Make a systems model of backward Euler method. CT block diagrams: adders, gains, and integrators:

X
y (t) = x(t) Backward Euler approximation: y [n] y [n 1] = x[n] T Equivalent system:

+ R

Backward Euler: substitute equivalent system for all integrators.

28

Model of Backward Euler Method

Replace every integrator in the CT system

X
with the backward Euler model:

+ R

Substitute the DT operator for A: 1 T T A= = 1 s 1R 1 z Backward Euler maps z 1 . 1 sT

29

Dependence of DT pole on Stepsize

Pole at z =
1 1+ T 1 . = 1 sT T

z = 0.1 z = 0.3 z =1 z
T

1 t 1 t 1 t 1 t 1 t
30

= 1.5 z

=2

Backward Euler: Mapping CT poles to DT poles

Backward Euler Map: s 0
1 T 2 T

1 z = 1 sT

1
1 2 1 3

s 0
1 z 1 sT

z 1 1

1 The entire left half-plane maps inside a circle with radius 1 2 at z = 2 .

If CT system is stable, then DT system is also stable.

31

Masses and Springs, Forwards and Backwards

In Homework 2, you investigated three numerical approximations to a mass and spring system: forward Euler backward Euler centered method

x ( t)

y (t)

32

Trapezoidal Rule
The trapezoidal rule uses centered dierences. Approximate CT signals at points between samples: 1 y [n] + yd [n 1] yc (n )T = d 2 2 Approximate derivatives at points between samples: 1 y [n] yd [n 1] y c (n )T = d 2 T
yc y c n 1 T 2 1 n T 2 y [n] + yd [n 1] = d 2 y [n] yd [n 1] = d T

yc (t) yd [n 1] (n 1)T
33

yd [n] t

nT

Trapezoidal Rule
The trapezoidal rule uses centered dierences. y (t) = x(t) Trapezoidal rule: y [n] y [n 1] x[n] + x[n 1] = T 2 Z transform: Y (s) T 1 + z 1 T H (z ) = = = 2 X (s) 2 1 z 1 Map: A= 1 T s 2 z+1 z1 . 1 sT 2 1 + sT 2

z+1 z1

Trapezoidal rule maps z

34

Trapezoidal Rule: Mapping CT poles to DT poles

Trapezoidal Map: s 0
1 T 2 T 1+ sT 2 1 sT 2

z=

1
1 3

0 1
2+jT 2jT

s 0
sT z 2+ 2sT

z 1 1

The entire left-half plane maps inside the unit circle. The j axis maps onto the unit circle
35

Mapping s to z: Leaky-Tank System

Forward Euler Method s

1 T 2 T

z 1

1 T

z 1 + sT

Backward Euler Method s

z 1

1 z 1 sT

Trapezoidal Rule s

z 1

sT z 2+ 2sT
36

Mapping s to z: Mass and Spring System

Forward Euler Method s

1 T 2 T

z 1

1 T

z 1 + sT

Backward Euler Method s

z 1

1 z 1 sT

Trapezoidal Rule s

z 1

sT z 2+ 2sT
37

Mapping s to z: Mass and Spring System

Forward Euler Method s

1 T 2 T

z 1

1 T

z 1 + sT

Backward Euler Method s

z 1

1 z 1 sT

Trapezoidal Rule s

z 1

sT z 2+ 2sT
38

Concept Map
Relations between CT and DT representations.
Delay R

Block Diagram
X +
Delay

System Functional
Y

+
Delay

Y 1 = H(R) = X 1 R R2

DT
index shift

Unit-Sample Response

h[n] : 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, . . .

T C X
Block Diagram
X +
1

Dierence Equation

y [n] = x[n] + y [n 1] + y [n 2]
AX

H (z ) =

System Function Y (z ) z2 = 2 X (z ) z z1

System Functional
Y

Y A = X A+

T D

Impulse Response
x (t) x(t)
1 et/ u(t) h(t) =

CT
Dierential Equation

r 1 (t) = r0 (t) r1 (t)

System Function Y (s) 1 H (s) = = X (s) 1 + s

39

MIT OpenCourseWare http://ocw.mit.edu

6.003 Signals and Systems

Fall 2011

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.