Upload
others
View
5
Download
0
Embed Size (px)
Citation preview
1
Discrete Time Systems
ELEC 3004: Systems: Signals & ControlsDr. Surya Singh(Some material adapted from courses by Russ Tedrake, MIT)
Lecture 15
[email protected]://robotics.itee.uq.edu.au/~elec3004/
© 2013 School of Information Technology and Electrical Engineering at The University of Queensland
April 26, 2012
• Properties of Discrete Time [DT] Signals
• DT Signal Models
• DT Signal Operations
• DT Convolution
• DT Systems Friday
• (also for Lab 3 / Exp 4): Introduce FIR Filters Friday
ELEC 3004: Systems 26 April 2013 - 2
Goals for the Week
2
Today…
ELEC 3004: Systems 26 April 2013 - 3
Week Date Lecture Title
127-FebIntroduction1-MarSystems Overview
26-MarSignals & Signal Models8-MarSystem Models
313-MarLinear Dynamical Systems15-MarSampling & Data Acquisition
420-MarTime Domain Analysis of Continuous Time Systems22-MarSystem Behaviour & Stability
527-MarSignal Representation29-MarHoliday
610-AprFrequency Response12-Aprz-Transform
717-AprNoise & Filtering19-AprAnalog Filters
824-AprDiscrete-Time Signals
26-AprDiscrete-Time Systems9
1-MayDigital Filters & IIR/FIR Systems3-MayFourier Transform & DTFT
108-MayState-Space
10-MayControllability & Observability
1115-MayIntroduction to Digital Control17-MayStability of Digital Systems
1222-MayPID & Computer Control24-MayInformation Theory & Communications
1329-MayApplications in Industry31-MaySummary and Course Review
Recap: AliasingAnother view of this
ELEC 3004: Systems 26 April 2013 - 4
3
• Aliasing - through sampling, two entirely different analogsinusoids take on the same “discrete time” identity
For f[k]=cosΩk, Ω=ωT:
The period has to be less than Fh (highest frequency):
Thus:ωf: aliased frequency:
ELEC 3004: Systems 26 April 2013 - 5
Recap: Alliasing
Basic Signal Processing: Continuous & Discrete
Continuous
Discrete
x(t)
h(t,T)
y(t,T)
x(n) y(n,k)
h(n,k) Delay
( )0
tdT
( )k
n
0
26 April 2013 -ELEC 3004: Systems 6
4
ELEC 3004: Systems 26 April 2013 - 7
Discrete LTI
• Circulant matrices are the “finite” equivalent of LTI systems.• Their properties are very similar to those of LTI systems.• The mathematics are a bit simpler, just “standard” linear algebra.• In particular, they have a nice and simple factorization structure.
ELEC 3004: Systems 26 April 2013 - 8
Discrete LTI Circulant Systems
5
• In general, matrices do not commute (AB ≠ BA)• Any two circulant matrices do commute!
ELEC 3004: Systems 26 April 2013 - 9
Circulant Systems
• A special circulant matrix
ELEC 3004: Systems 26 April 2013 - 10
Shift matrix Z
6
• The shift Z acts on signals like a unit delay (but it “wraps around”)
• Recall time-invariance: “shifting the input shifts the output”
• A natural notion: linear shift-invariant systems
or equivalently, DZ = ZD (“system commutes with the shift”)
ELEC 3004: Systems 26 April 2013 - 11
Shift invariant systems
ELEC 3004: Systems 26 April 2013 - 12
Shift invariant systems are circulants
7
• Eigenvalues (λ) and Characteristic Functions:
• For DT LTI the eigenfunctions are complex exponentials
ELEC 3004: Systems 26 April 2013 - 13
Eigenvalues of DT-LTI Systems
• Every circulant matrix has an eigenvalue decomposition:
eigenvectors:• Notice that the eigenvectors T do not depend on entries ai!• This change-of-base matrix T is called the Fourier matrix.
ELEC 3004: Systems 26 April 2013 - 14
Back to Circulant Matrices…Their eigenvectors do not depend on ai
8
• They do depend on the values of ai (a0, a1, …, an-1).(you’d hope)
ELEC 3004: Systems 26 April 2013 - 15
Eigenvalues of Circulant Matrices
ELEC 3004: Systems 26 April 2013 - 16
Ex:
9
• Ex: Kinematic (constantly accelerating body) systems:
• Discrete Time Convolution in Matrix Formvia the Toeplitz Matrix
ELEC 3004: Systems 26 April 2013 - 17
WHY?:These Systems Show Up Everywhere
• Remember: Flip Shift Slide SumDT Convolution: Vector Sums
x(n) = 1 2 3 4 5 h(n) = 3 2 1
0 0 1 2 3 4 5
1 2 3 0 0 0 0
0 0 1 2 3 4 5
0 1 2 3 0 0 0
0 0 1 2 3 4 5
0 0 1 2 3 0 0
x(k)
h(n,k)
3 2 6 1 4 9y(n,k)
e.g. convolution
y(n) 3 8 14
Sum over all k
Notice the gain
h(n-k)
26 April 2013 -ELEC 3004: Systems 18
10
DT Convolution: Matrix Formulation of Convolution
3
8
14
20
26
14
5
1 2 3 0 0 0 0
0 1 2 3 0 0 0
0 0 1 2 3 0 0
0 0 0 1 2 3 0
0 0 0 0 1 2 3
0 0 0 0 0 1 2
0 0 0 0 0 0 1
0 0
0 0
0 0
0 0
0 0
3 0
2 3
0
0
1
2
3
4
5
0
0
.
y H x
Toeplitz Matrix
26 April 2013 -ELEC 3004: Systems 19
• Impulse Response:Imaging a discrete-time, LTI System F . It’s impulse response is given by: • Arbitrary Response:An arbitrary input x[n] can be written: (Sampling theorem)
So:
∴ a DT LTI is completely characterized by its impulse responseELEC 3004: Systems 26 April 2013 - 20
Response of a Discrete-Time LTI System & Convolution Sum
11
• As with the continuous domain, commutatively, associativity, and distributivity hold…
• Commutativity gives a nice result:
• Makes Step Responses “Easy”:
ELEC 3004: Systems 26 April 2013 - 21
Convolution Response
• Causality:
» or
• Input is Causal if:
• Then output is Causal:
• And, DT LTI is BIBO stable if:
ELEC 3004: Systems 26 April 2013 - 22
DT Causality & BIBO Stability
12
ELEC 3004: Systems 26 April 2013 - 23
Impulse Response (Graphically)
∞ matrix × ∞ vector?
• First let’s multiply circulant matrices…– A circulant matrix can be descibed completely by its first row or column
• Multiply by u[k]
∴ For circulant matrices, matrix multiplication reduces to a weighted combination of shifted impulse responses
ELEC 3004: Systems 26 April 2013 - 24
How do you multiply an infinite matrix?
Z: Shift operator
13
Two Types of Systems• Linear shift-invariant:
Z: Shift operator
• Linear time-invariant system
R: Unit delay operator
ELEC 3004: Systems 26 April 2013 - 25
Impulse Response of Both Types
ELEC 3004: Systems 26 April 2013 - 26
14
Impulse Response of Both Types
ELEC 3004: Systems 26 April 2013 - 27
Next Time in Linear Systems ….
ELEC 3004: Systems 26 April 2013 - 28
Week Date Lecture Title
127-FebIntroduction1-MarSystems Overview
26-MarSignals & Signal Models8-MarSystem Models
313-MarLinear Dynamical Systems15-MarSampling & Data Acquisition
420-MarTime Domain Analysis of Continuous Time Systems22-MarSystem Behaviour & Stability
527-MarSignal Representation29-MarHoliday
610-AprFrequency Response12-Aprz-Transform
717-AprNoise & Filtering19-AprAnalog Filters
824-AprDiscrete-Time Signals
26-AprDiscrete-Time Systems9
1-MayDigital Filters & IIR/FIR Systems3-MayFourier Transform & DTFT
108-MayState-Space
10-MayControllability & Observability
1115-MayIntroduction to Digital Control17-MayStability of Digital Systems
1222-MayPID & Computer Control24-MayInformation Theory & Communications
1329-MayApplications in Industry31-MaySummary and Course Review