MIT 6.003 signals & systems video note

21:55 2013-5-29 Wednesday
mit 6.003(signals & systems)
the 6.003 abstraction(input -> system -> output)
22:12 2013-5-29
ZOH == Zero Order Holder
//
20:41 2013-5-30 Thursday
difference equation
CT, DT...
unit sample
20:46 2013-5-30
block diagram: hardware realization...
20:55 2013-5-30
from samples to signals...
R: right shift operator
20:57 2013-5-30
"operator" acts on "signals"
initially at rest....
21:14 2013-5-30
"recipe" <-> "constraint"
imperative
synthetic division
21:17 2013-5-30
"feedback" means there is a cycle somewhere
acyclic
21:20 2013-5-30
signal cyclic signal path...
21:44 2013-5-30
"initially at rest" means all output of "delay" being ZERO!
21:52 2013-5-30
next level of abstraction...
23:19 2013-5-30
What's special about feedback?
transient input -> persistent output(echo?)
23:20 2013-5-30
cyclic signal paths
23:29 2013-5-30
unit sample response
unit step response
23:30 2013-5-30
geometric growth
23:32 2013-5-30
second order systems
R = Right shift(Delay)
///
21:44 2013-5-31 Friday
target: weekend video 3 ~ 18(16 videos)
unit sample response
unit step response
divergent <-> convergent
22:14 2013-5-31
cyclic system(with feedback)
22:18 2013-5-31
feedback system -> feedforward system
22:52 2013-5-31
R = 1/z
/
17:25 2013-6-1 Saturday
fundamental mode
complex poles
18:08 2013-6-1
A operator == Accumulate(integrate)
18:16 2013-6-1
watching Freeman's 6.003 video 4
CT system........
DT system: R == Right Shift
CT system: A == Accumulate
18:30 2013-6-1
unit sample(DT)
unit impulse(CT)
18:47 2013-6-1
differential equations
complex exponential
18:49 2013-6-1
How to solve CT Feedback system?
1. differential equation
2. operator method
18:57 2013-6-1
expand in ascending series....
19:03 2013-6-1
basis function....
19:07 2013-6-1
convergent <-> divergent
19:10 2013-6-1
unit circle.....
19:17 2013-6-1
replace:
D -> 1/z
A -> 1/s
19:20 2013-6-1
Taylor's series expansion
22:40 2013-6-1
What's the relation of
unit sample response <-> system functional
22:53 2013-6-1
system functional H(R)
unit sample response h(n)
22:56 2013-6-1
Z-transform
bilateral z-transform
unilateral z-transform
23:07 2013-6-1
ROC == Region Of Convergence
23:12 2013-6-1
adders, delays, gains.......
23:19 2013-6-1
ratio of polynomials
23:21 2013-6-1
fundamental theorem of algebra
23:36 2013-6-1
the switch of region(ROC)
each term is getting more convergent...
23:38 2013-6-1
right-sided signal -> outside region
23:41 2013-6-1
iterating backwards in time
23:46 2013-6-1
Freeman 6.003 lec4, 5 done!
review them tomorrow!
///
11:10 2013-6-2 Sunday
Laplace transform
bilateral Laplace transform
unilateral Laplace transform
11:41 2013-6-2
the region is alwayls delimited by poles....
11:49 2013-6-2
Laplace transform can be viewed as a kind of Fourier
transform after modulated by exp(-σt)...
12:07 2013-6-2
Why bother Laplace transform?
differential equations -> algebra equations
12:09 2013-6-2
the Laplace transform of sum is
sum of Laplace transform........
12:11 2013-6-2
integrate by part....
12:16 2013-6-2
sifting property
12:17 2013-6-2
differential equation -> algebraic equation
12:25 2013-6-2
IVT == Initial Value Theorem
Final Value Theorem
12:26 2013-6-2
system functional H(A), system function H(s)
12:34 2013-6-2
leaky tank system
12:34 2013-6-2
DT system, CT system....
12:36 2013-6-2
Discrete Approximation of CT system...
12:38 2013-6-2
step response of the leaky tank system?
12:48 2013-6-2
unit impulse response: δ(t)
unit step response: u(t)
12:51 2013-6-2
Forward Euler method
Forward Euler Approximation........
12:55 2013-6-2
Approximate derivatives.......
12:55 2013-6-2
forward Euler
13:17 2013-6-2
Z = 1 + sT
13:24 2013-6-2
CT pole -> DT pole
13:24 2013-6-2
backward Euler Approximation......
13:25 2013-6-2
A == Accumulate, (A == 1/s)
13:29 2013-6-2
BE == Backward Euler
FE == Forward Euler
13:30 2013-6-2
Backward Euler:
mapping CT poles -> DT poles....
13:33 2013-6-2
BE(Backward Euler) can guarantee that
stable CT system -> stable DT system....
13:35 2013-6-2
Trapezoidal method
13:46 2013-6-2
mapping s to z
15:00 2013-6-2
unit sample response: h[n]
weighted shifted unit sample......
15:01 2013-6-2
LTI -> convolution......
15:04 2013-6-2
"signals", not "samples"
15:10 2013-6-2
the sequence of convolution:
1. flip 2. shift 3. multiply 4. sum
15:24 2013-6-2
unit sample response
15:26 2013-6-2
CT convolution
15:47 2013-6-2
computational tool
15:54 2013-6-2
spherical wave........
15:55 2013-6-2
PSF == Point Spread Function(3D impulse response)
16:25 2013-6-2
Hubble Space Telescope
blur, deblur, convolution.......
16:32 2013-6-2
Frequency response
unit samle response: h(n)
unit impulse response: h(t)
16:40 2013-6-2
LTI == Linear Time-Invariant
16:44 2013-6-2
How to characterize a system?
1. unit sample response: h(n)
unit impulse response: h(t)
2. Frequency response......
17:07 2013-6-2
out of phase...
17:11 2013-6-2
How to find frequency response?
method 1: differential equations
method 2: impulse response
method 3: eigenfunctions, eigenvalues
17:29 2013-6-2
complex exponential
17:36 2013-6-2
complex exponentials are eigenfunctions of LTI system......
17:41 2013-6-2
system function H(s) is the "eigenvalue" of LTI system, when
"eigenfunction" is exp(st)
17:42 2013-6-2
fundamental theorem of algebra
17:42 2013-6-2
LCCDE == Linear Constant Coefficient Differential Equation
17:50 2013-6-2
complex conjugate......
17:52 2013-6-2
sinusoidal signal is not the "eigenfunction" of LTI systems!
17:52 2013-6-2
Frequency response:
response to eternal sinusoids....
17:55 2013-6-2
pole-zero diagram
17:55 2013-6-2
vector diagram
18:01 2013-6-2
magnitude, phase....
18:45 2013-6-2
negative frequency
/
21:16 2013-6-3 Monday
start 6.003 Freeman lec 10,
Feedback & control
21:23 2013-6-3
lift coefficient, drag coefficient......
21:24 2013-6-3
wind tunnel
flow visualization........
21:26 2013-6-3
open loop <-> feedback.
21:29 2013-6-3
structure of a control system:
controller + plant + sensor
21:29 2013-6-3
reference command
21:33 2013-6-3
unit sample response: h(n)
unit step response: s(n)
21:34 2013-6-3
fastest convergence
speed of convergence.........
21:41 2013-6-3
feedback path....
21:44 2013-6-3
How to find poles in "system functional"?
1. replace R -> 1/s in system functional
2. resolve for poles in H(s)
21:47 2013-6-3
fast repsonse, slow response....
21:47 2013-6-3
complex poles
21:47 2013-6-3
root locus!
21:51 2013-6-3
change K to tune the response
21:51 2013-6-3
closed-loop poles
21:52 2013-6-3
convergence, convergence rate
21:52 2013-6-3
the biggest pole dominates...
21:53 2013-6-3
in generally, delay is a bad thing
21:54 2013-6-3
proportional feedback
22:10 2013-6-3
bode plot
22:10 2013-6-3
pole-zero diagram
22:11 2013-6-3
small number of singularities
22:14 2013-6-3
complex exponential is the eigenfunction of LTI systems
22:15 2013-6-3
vector diagram
22:16 2013-6-3
Frequency response...
22:20 2013-6-3
magnitude response
phase response.......
22:22 2013-6-3
log-log axes
22:22 2013-6-3
asymptotic behavior
22:24 2013-6-3
a good approximation
22:27 2013-6-3
cirtical frequency
22:35 2013-6-3
Why use negative frequencies?
because we like Euler's formula
, with Euler we represent sine & cosine
as complex exponential which is the eigenfunction
of LTI systems.......
22:38 2013-6-3
low frequency asymptote
high frequency asymprote.........
22:52 2013-6-3
Bode plot: adding instead of multiplying
22:54 2013-6-3
angle frequency
22:55 2013-6-3
straight line approximation
22:59 2013-6-3
Bode plot turn something conceptually hard to sth
simple(just consider poles & zeros)
23:02 2013-6-3
Bode approximation
23:06 2013-6-3
log scale.......
23:07 2013-6-3
frequency in log scale
dB == 20 * log (A)
23:09 2013-6-3
unit of frequency is decade.........
23:14 2013-6-3
mass-spring-dashpot system
23:14 2013-6-3
complex poles......
23:15 2013-6-3
complex conjugate
23:15 2013-6-3
complex pairs(complex conjugate)
23:18 2013-6-3
High-Q system
23:21 2013-6-3
crossover
23:22 2013-6-3
isolated zeros, isolated poles....
23:30 2013-6-3
root locus
23:32 2013-6-3
to enhance performance using feedback
23:33 2013-6-3
ideal op-amps
23:33 2013-6-3
broad bandwidth
23:34 2013-6-3
LM741
23:35 2013-6-3
audio signal processing
23:36 2013-6-3
fast time response
23:37 2013-6-3
unit step response
23:46 2013-6-3
How to find τ of unit step response of LM741,
assuming know frequency response diagram?
frequency response -> poles, zeros -> H(s) -> h(t) -> s(t)
23:49 2013-6-3
circuit diagram -> 6.003 model
23:49 2013-6-3
dorminant pole
///
19:53 2013-6-4 Tuesday
start 6.003 Freeman lec 12!
20:19 2013-6-4
How to find time constant of s(t) from frequency response?
frequency response(Bode plot) -> pole
->H(s) -> h(t) -> s(t)
20:31 2013-6-4
dorminant pole
20:42 2013-6-4
trade gain for sth else(speed, bandwidth etc.)
20:51 2013-6-4
velocity controller <-> position controller
21:03 2013-6-4
sluggish step -> fast step
21:15 2013-6-4
voltage controlled velocity
voltage controlled position
21:26 2013-6-4
power amplifier
21:37 2013-6-4
crossover distortion
21:41 2013-6-4
feedback to improve crossover distortion.....
21:48 2013-6-4
magnetic levitation
22:14 2013-6-4
equilibrium
22:34 2013-6-4
DC means zero frequency
22:34 2013-6-4
DC, fundamental, harmonics
22:38 2013-6-4
harmonic structure
22:42 2013-6-4
periodic signal
22:56 2013-6-4
harmonic component
22:56 2013-6-4
analysis equation <-> synthesis equation
23:28 2013-6-4
energy difference
23:29 2013-6-4
signal with finite energy
23:31 2013-6-4
complex exponential: eigenfunction of LTI systems
23:39 2013-6-4
harmonic decomposition
23:40 2013-6-4
system as a filter
///
19:59 2013-6-5 Wednesday
Fourier Series -> orthogonal decomposition
20:02 2013-6-5
dot product -> inner product.....
20:11 2013-6-5
inner product depends on period
//
15:26 2013-6-11 Tuesday
mit 6.003 Freeman
Lec 20 done!
/
17:44 2013-6-12 Wednesday
mit 6.003 Freeman all video done!