Modeling and Analyzing System Behavior: February 25, 2013

6.
01: Introduction to EECS I

Modeling and Analyzing System Behavior
February 25, 2013
6.01: Introduction to EECS I

The intellectual themes in 6.01 are recurring themes in EECS: design of complex systems modeling and controlling physical systems augmenting physical systems with computation building systems that are robust to uncertainty
Intellectual themes are developed in context of a mobile robot.
Goal is to convey a distinct perspective about engineering.
Modeling and Analyzing Behavior

Consider the behavior that results if you program the robot to move forward or backward so that it is half a meter from the wall in front of it (Design Lab 1).
desiredDistance currentDistance
Demo: consider two possible wall nder behaviors.
Check Yourself
Which behavior is better? 1. 2. 3. 4. Behavior Behavior Both are Both are 1 2 good bad
Check Yourself
Which behavior is better? Behavior 1 seems unneccessarily sluggish. Behavior 2 overshoots. Performance metrics: How long does it take for the output signal to reach its nal value? Does the output signal overshoot? If so, how much?
How can we optimize performance? What are the fundamental limitations?
Modeling and Analyzing Behavior

Goal: develop a framework for analyzing behavior. Three parts computational model state machines mathematical model dierence equations analysis signals and systems
State Machines
State machines are convenient abstractions for organizing computations that evolve with time.
input in
state sn
output on
On the nth step, the system gets input in generates output on and moves to a new state sn+1 Output and next state depend on input and current state
Explicit representation of stepwise nature of required computation.
State Machines
Example: accumulator On each step, an input is added to the already accumulated inputs, and the result becomes the new output.
Accumulator
Assume state is initially zero.
s0 = 0
Accumulator
input i0 = 1
s0 = 0 s1 = 1
output o0 = 1
Accumulator
input
s1 = 1
output
Accumulator
input i1 = 2
s1 = 1 s2 = 3
output o1 = 3
Accumulator
input
s2 = 3
output
Accumulator
input i2 = 3
s2 = 3 s3 = 6
output o2 = 6
Accumulator
input
s3 = 6
output
Accumulator
input i3 = 4
s3 = 6 s4 = 10
output o3 = 10
Accumulator
input
s4 = 10
output
Accumulator
input i4 = 5
s4 = 10 s5 = 15
output o4 = 15
State Machines in Python

Using OOP to represent state machines. SM Class: All state machines share some methods: start() initialize the state machine step(inp) receive and process new input transduce(inputs) make repeated calls to step Accumulator Subclass of SM: All accumulators share startState initial contents of state getNextValues(state, inp) method to process input Instances of Accumulator: Each instance has its own private state state sum of inputs so far
SM Class
The generic methods of the SM class use startState to initialize the instance variable state. Then getNextValues is used to process inputs, so that step can update state.
class SM: startState= None def getStartState(self): return self.startState def start(self): self.state = self.getStartState() def step(self, inp): (s, o) = self.getNextValues(self.state, inp) self.state = s return o def transduce(self, inputs): self.start() return [self.step(inp) for inp in inputs]
Note that getNextValues should not change state. The state is managed by start and step.
Accumulator
class Accumulator(SM): startState = 0 def getNextValues(self, state, inp): return (state + inp, state + inp)
Check Yourself
>>> >>> >>> >>> >>> >>> >>> >>>
a = Accumulator() a.start() a.step(7) b = Accumulator() b.start() b.step(10) a.step(-2) print a.state,a.getNextValues(8,13),b.getNextValues(8,13)
What will be printed? 1: 2: 3: 4: 5:
5 (18, 18) (23, 23) 5 (21, 21) (21, 21) 15 (18, 18) (23, 23) 15 (21, 21) (21, 21)
none of the above
Environments Associated with Accumulator

a = Accumulator() a.start() a.step(7) b = Accumulator() b.start() b.step(10) a.step(-2)
SM Accumulator a b startState getStartState start step transduce startState getNextValues state state None (proc) (proc) (proc) (proc) 0 (proc) 5 10
Check Yourself
>>> >>> >>> >>> >>> >>> >>> >>>
a = Accumulator() a.start() a.step(7) b = Accumulator() b.start() b.step(10) a.step(-2) print a.state,a.getNextValues(8,13),b.getNextValues(8,13)
What will be printed? 1: 2: 3: 4: 5:
5 (18, 18) (23, 23) 5 (21, 21) (21, 21) 15 (18, 18) (23, 23) 15 (21, 21) (21, 21)
none of the above
State Machines
The state machine representation for controlling processes is simple and concise separates system specication from looping structures over time is modular We will use this approach in controlling our robots and in simulating the resulting behavior.
Analyzing System Behavior

We will analyze behavior by examining properties of output signals.
signal in
system
signal out
Demo: consider the wall nder behavior.
Check Yourself
Which plot best represents currentDistance?
1. 2.
3.
4.
5. none of the above
Check Yourself
Which plot best represents currentDistance?
2.
1. 2.
3.
4.
5. none of the above
Performance Analysis
Quantify performance by characterizing input and output signals.
desiredDistance
n
wallFinder
system
currentDistance
n
The Signals and Systems Abstraction

Represent a system (physical, mathematical, or computational) by the way it transforms an input signal into an output signal.
signal in
system
signal out
Look at representations that facilitate analysis
Dierence Equations
We can represent the transformation from input to output signal by a dierence equation. Dierence equations are mathematically precise and compact. Example: y [n] = x[n] x[n 1] Let x[n] equal the unit sample signal [n], [ n] = 1, 0, if n = 0; otherwise. x[ n] = [ n]
n 1 0 1 2 3 4
Dierence Equations
Convenient for step-by-step analysis.
Find y [n] given x[n] = [n]:
y [n] = x[n] x[n 1] y [1] = x[1] x[2] y [0] = x[0] x[1] y [1] = x[1] x[0] y [2] = x[2] x[1] y [3] = x[3] x[2] ... =00=0 =10=1 = 0 1 = 1 =00=0 =00=0
x[ n] = [ n]
y [ n]
n 1 0 1 2 3 4 1 0 1 2 3 4
Dierence Equations
x[ n] = [ n]
y [ n]
n 1 0 1 2 3 4 1 0 1 2 3 4
Dierence Equations
x[ n] = [ n]
y [ n]
n 1 0 1 2 3 4 1 0 1 2 3 4
Dierence Equations
x[ n] = [ n]
y [ n]
n 1 0 1 2 3 4 1 0 1 2 3 4
Dierence Equations
x[ n] = [ n]
y [ n]
n 1 0 1 2 3 4 1 0 1 2 3 4
Dierence Equations
x[ n] = [ n]
y [ n]
n 1 0 1 2 3 4 1 0 1 2 3 4
Dierence Equations
x[ n] = [ n]
y [ n]
n 1 0 1 2 3 4 1 0 1 2 3 4
Block Diagrams
Graphic representation of the dierence equation. Represent y [n] = x[n] x[n 1] with a block diagram:
x[ n ] 1
Delay
y [ n]
x[ n] = [ n]
y [ n]
n 1 0 1 2 3 4 1 0 1 2 3 4
Step-By-Step Solutions
Block diagrams are also useful for step-by-step analysis. Represent y [n] = x[n] x[n 1] with a block diagram: start at rest
x[ n ] 1
Delay
+ 0
y [ n]
x[ n] = [ n]
y [ n]
n 1 0 1 2 3 4 1 0 1 2 3 4
1 1 1
Delay
+ 0
x[ n] = [ n]
y [ n]
n 1 0 1 2 3 4 1 0 1 2 3 4
10 1 1
Delay
+ 0 1
x[ n] = [ n]
y [ n]
n 1 0 1 2 3 4 1 0 1 2 3 4
10 1 0
Delay
1 0 1
x[ n] = [ n]
y [ n]
n 1 0 1 2 3 4 1 0 1 2 3 4
0 1 0
Delay
+ 1
x[ n] = [ n]
y [ n]
n 1 0 1 2 3 4 1 0 1 2 3 4
0 1 0
Delay
1 1 0
x[ n] = [ n]
y [ n]
n 1 0 1 2 3 4 1 0 1 2 3 4
0 1 0
Delay
0 1 0
x[ n] = [ n]
y [ n]
n 1 0 1 2 3 4 1 0 1 2 3 4
0 1 0
Delay
+ 0
x[ n] = [ n]
y [ n]
n 1 0 1 2 3 4 1 0 1 2 3 4
0 1 0
Delay
+ 0
x[ n] = [ n]
y [ n]
n 1 0 1 2 3 4 1 0 1 2 3 4
0 1 0
Delay
+ 0
x[ n] = [ n]
y [ n]
n 1 0 1 2 3 4 1 0 1 2 3 4
Multiple Representations of Discrete-Time Systems

Block diagrams are useful alternative representations that highlight visual/graphical patterns. Dierence equation: y [n] = x[n] x[n 1] Block diagram: x[ n]
+ 1
Delay
y [ n]
Same input-output behavior, dierent strengths/weaknesses: dierence equations are mathematically compact block diagrams illustrate signal ow paths
From Samples to Signals

Lumping all of the (possibly innite) samples into a single object the signal simplies its manipulation. This lumping is analogous to representing coordinates in three-space as points representing lists of numbers as vectors in linear algebra creating an object in Python
From Samples to Signals

Operators manipulate signals rather than individual samples.
X 1
Delay
Nodes represent whole signals (e.g., X and Y ). The boxes operate on those signals: Delay = shift whole signal to right 1 time step Add = sum two signals 1: multiply by 1
Signals are the primitives. Operators are the means of combination.
Operator Notation
Symbols can now compactly represent diagrams. Let R represent the right-shift operator: Y = R{X } RX where X represents the whole input signal (x[n] for all n) and Y represents the whole output signal (y [n] for all n)
Operator Notation
Systems are concisely represented with operators. Dierence machine:
X 1
Delay
Equivalent representation with R: Y = X RX = (1 R) X
Operator Notation: Check Yourself
Let Y = RX . Which of the following is/are true:
1. y [n] = x[n] for all n 2. y [n + 1] = x[n] for all n 3. y [n] = x[n + 1] for all n 4. y [n 1] = x[n] for all n 5. none of the above
Operator Notation: Check Yourself
Let Y = RX . Which of the following is/are true:
2.
1. y [n] = x[n] for all n 2. y [n + 1] = x[n] for all n 3. y [n] = x[n + 1] for all n 4. y [n 1] = x[n] for all n 5. none of the above
Operator Representation of a Cascaded System

System operations have simple operator representations. Cascade systems multiply operator expressions.
X 1
Delay
Y1
+ 1
Delay
Y2
Using operator notation: Y1 = (1 R) X Y2 = (1 R) Y1 Substituting for Y1 : Y2 = (1 R)(1 R) X
Operator Algebra
Operator expressions expand and reduce like polynomials.
X 1
Delay
Y1
+ 1
Delay
Y2
Using dierence equations: y2 [n] = y1 [n] y1 [n 1] = (x[n] x[n 1]) (x[n 1] x[n 2]) = x[n] 2x[n 1] + x[n 2] Using operator notation: Y2 = (1 R) Y1 = (1 R)(1 R) X = (1 R)2 X = (1 2R + R2 ) X
Operator Approach
Applies your existing expertise with polynomials to understand block diagrams, and thereby understand systems.
Operator Algebra
Operator notation facilitates seeing relations among systems. Equivalent block diagrams (assuming both initially at rest):
X 1
Delay
Y1
+ 1
Delay
Y2
X
Delay
+ 2
Delay
Equivalent operator expression: (1 R)(1 R) = 1 2R + R2
Operator Algebra
Operator notation prescribes operations on signals, not samples: e.g., start with X , subtract 2 times a right-shifted version of X , and add a double-right-shifted version of X !
X:
n 1 0 1 2 3 4 5 6
2RX :
n 1 0 1 2 3 4 5 6
+ R2 X :
n 1 0 1 2 3 4 5 6
y = X 2 RX + R2 X :
n 1 0 1 2 3 4 5 6
Operator Algebra
Expressions involving R obey many familiar laws of algebra, e.g., commutativity. R(1 R) X = (1 R)RX This is easily proved by the denition of R, and it implies that cascaded systems commute (assuming initial rest)
X 1
is equivalent to Delay
Delay
Delay
+ 1
Delay
Operator Algebra
Multiplication distributes over addition. Equivalent systems
X 1
Delay
Delay
X 1
Delay Delay Delay
Equivalent operator expression: R(1 R) = R R2
Operator Algebra
The associative property similarly holds for operator expressions. Equivalent systems
X 1
Delay Delay Delay
+ 1
2
Delay
X 1
Delay
+ 1
2
Delay
Delay Delay
Equivalent operator expression: (2 R) R(1 R) = (2 R)R (1 R)
Check Yourself
How many of the following systems are equivalent?
Delay
Delay
Delay
Delay
Delay
Delay
Check Yourself
How many of the following systems are equivalent?
Delay
Delay
Delay
Delay
Delay
Delay
Feedback
Feedback complicates relation between input and output signals.
X 1
Delay
Y
Y = (1 R) X
Without feedback, output signal is linear combination of shifted versions of input signal.
+
Delay
Y
Y = RY + X (1 R) Y = X
Feedback introduces a similar constraint, but now the input signal is a linear combination of shifted versions of the output signal. But how do we nd Y ?
Example: Accumulator
Try step-by-step analysis: it always works. Start at rest.
x[ n]
+
Delay
y [ n]
y [n] = x[n] + y [n 1] y [0] = x[0] + y [1] = 1 + 0 = 1 y [1] = x[1] + y [0] y [2] = x[2] + y [1] ... y [ n] =0+1=1 =0+1=1
x[ n] = [ n]
n 1 0 1 2 3 4
Persistent response to a transient input!
n 1 0 1 2 3 4
x[ n]
+
Delay
y [ n]
x[ n] = [ n]
n 1 0 1 2 3 4
n 1 0 1 2 3 4
x[ n]
+
Delay
y [ n]
x[ n] = [ n]
n 1 0 1 2 3 4
n 1 0 1 2 3 4
x[ n]
+
Delay
y [ n]
x[ n] = [ n]
n 1 0 1 2 3 4
n 1 0 1 2 3 4
x[ n]
+
Delay
y [ n]
x[ n] = [ n]
n 1 0 1 2 3 4
n 1 0 1 2 3 4
x[ n]
+
Delay
y [ n]
x[ n] = [ n]
n 1 0 1 2 3 4
n 1 0 1 2 3 4
The response of the accumulator system could also be generated by a system with innitely many paths from input to output, each with one unit of delay more than the previous.
X
Delay Delay Delay Delay Delay Delay
...
Y = (1 + R + R2 + R3 + ) X
...
These systems are equivalent in the sense that if each is initially at rest, they will produce identical outputs from the same input. Y2 = (1 + R + R2 + R3 + ) X2
(1 R) Y1 = X1
Proof: Assume X2 = X1 : Y2 = (1 + R + R2 + R3 + ) X2 = (1 + R + R2 + R3 + ) X1 = (1 + R + R2 + R3 + ) (1 R) Y1 = ((1 + R + R2 + R3 + ) (R + R2 + R3 + )) Y1 = Y1 It follows that Y2 = Y1 .
The system functional for the accumulator is the reciprocal of a polynomial in R. + X Y Delay (1 R) Y = X
The product (1 R) (1 + R + R2 + R3 + ) equals 1. Therefore the terms (1 R) and (1 + R + R2 + R3 + ) are reciprocals. Thus we can write Y 1 = = 1 + R + R2 + R3 + R4 + X 1R
The reciprocal of 1 R can also be evaluated using synthetic division.
1 +R +R2 +R3 + 1R 1 1 R R 2 R2 R2 3 R3 R3 4
Therefore 1 = 1 + R + R2 + R3 + R4 + 1R
Check Yourself
A system is described by the following operator expression: Y 1 = . X 1 + 2R Determine the output of the system when the input is a unit sample.
Check Yourself
Evaluate the system function using synthetic division.
1 2R +4R2 8R3 + 1 + 2R 1 1 +2R 2R 2R 4R2 4R2 4R2 +8R3 8R3 8R3 16R4
Therefore the system function can be written as Y 1 = = 1 2R + 4R2 8R3 + 16R4 + X 1 + 2R
Check Yourself
Now nd Y given that X is a delta function. x[ n] = [ n] Think about the sample representation of the system function: Y = 1 2R + 4R2 8R3 + 16R4 + X y [n] = 1 2R + 4R2 8R3 + 16R4 + [n] y [n] = [n] 2 [n 1] + 4 [n 2] 8 [n 3] + 16 [n 4] +
Check Yourself
A system is described by the following operator expression: 1 Y = . X 1 + 2R Determine the output of the system when the input is a unit sample.
y [n] = [n] 2 [n 1] + 4 [n 2] 8 [n 3] + 16 [n 4] +
y [ n]
Linear Dierence Equations with Constant Coecients

Any system composed of adders, gains, and delays can be represented by a dierence equation. y [n] + a1 y [n 1] + a2 y [n 2] + a3 y [n 3] + = b0 x[n] + b1 x[n 1] + b2 x[n 2] + b3 x[n 3] + Such a system can also be represented by an operator expression. (1 + a1 R + a2 R2 + a3 R3 + ) Y = (b0 + b1 R + b2 R2 + b3 R3 + ) X We will see that this correspondence provides insight into behavior. This correspondence also reduces algebraic tedium.
Check Yourself
Determine the dierence equation that relates x[] and y [].
x[ n]
Delay Delay
y [ n]
1. 2. 3. 4. 5.
y [n] = x[n 1] + y [n 1] y [n] = x[n 1] + y [n 2] y [n] = x[n 1] + y [n 1] + y [n 2] y [n] = x[n 1] + y [n 1] y [n 2] none of the above
Check Yourself
Determine a dierence equation that relates x[] and y [] below.
x[ n]
Delay Delay
y [ n]
Assign names to all signals. Replace Delay with R.
E W
R R
Express relations among signals algebraically. E =X +W ; Solve: Y = RE ; W = RY RX = Y R2 Y
Y = RE = R(X + W ) = R(X + RY )
Corresponding dierence equation:
y [n] = x[n 1] + y [n 2]
Check Yourself
Determine the dierence equation that relates x[] and y []. 2.
x[ n]
Delay Delay
y [ n]
1. 2. 3. 4. 5.
y [n] = x[n 1] + y [n 1] y [n] = x[n 1] + y [n 2] y [n] = x[n 1] + y [n 1] + y [n 2] y [n] = x[n 1] + y [n 1] y [n 2] none of the above
Multiple Representations of Systems

State machines: computationally ecient. input in state sn output on
Dierence equations: mathematically compact. y [n] = x[n] x[n 1] Block diagrams: illustrate signal ow paths.
x[ n] 1
Delay
y [ n]
Operator representations: analyze systems as polynomials. Y = (1 R) X
State Machine Combinators

State machines can be combined for more complicated tasks.
Cascade(M1,M2) input i1 M1 o1=i2 M2 o2 output
Parallel(M1,M2) input i1 i2 M1 M2 o1 o2 output[0] input[0] (tuple) (tuple) output[1] input[1]
Parallel(M1,M2) i1 i2 M1 M2 o1 o2 output[0] (tuple) output[1]
Feedback(M1) i1 M1 o1 output input
Feedback2(M1) i1[0] o1 i1[1] M1
output
This Week
Software lab: Representing systems with state machines. Design lab: Modeling the wall nder behavior.

Modeling and Analyzing System Behavior: February 25, 2013

Uploaded by

Copyright:

Available Formats

Modeling and Analyzing System Behavior: February 25, 2013

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Modeling and Analyzing System Behavior: February 25, 2013

Uploaded by

Copyright:

Available Formats

6.

01: Introduction to EECS I

February 25, 2013

6.01: Introduction to EECS I

Intellectual themes are developed in context of a mobile robot.

Goal is to convey a distinct perspective about engineering.

Modeling and Analyzing Behavior

How can we optimize performance? What are the fundamental limitations?

Modeling and Analyzing Behavior

Explicit representation of stepwise nature of required computation.

State Machines in Python

>>> >>> >>> >>> >>> >>> >>> >>>

a = Accumulator() a.start() a.step(7) b = Accumulator() b.start() b.step(10) a.step(-2) print a.state,a.getNextValues(8,13),b.getNextValues(8,13)

What will be printed? 1: 2: 3: 4: 5:

Environments Associated with Accumulator

>>> >>> >>> >>> >>> >>> >>> >>>

a = Accumulator() a.start() a.step(7) b = Accumulator() b.start() b.step(10) a.step(-2) print a.state,a.getNextValues(8,13),b.getNextValues(8,13)

What will be printed? 1: 2: 3: 4: 5:

Analyzing System Behavior

Demo: consider the wall nder behavior.

Which plot best represents currentDistance?

5. none of the above

Which plot best represents currentDistance?

5. none of the above

The Signals and Systems Abstraction

Look at representations that facilitate analysis

Find y [n] given x[n] = [n]:

Find y [n] given x[n] = [n]:

Find y [n] given x[n] = [n]:

Find y [n] given x[n] = [n]:

Find y [n] given x[n] = [n]:

Find y [n] given x[n] = [n]:

Find y [n] given x[n] = [n]:

Multiple Representations of Discrete-Time Systems

From Samples to Signals

From Samples to Signals

Signals are the primitives. Operators are the means of combination.

Equivalent representation with R: Y = X RX = (1 R) X

Operator Notation: Check Yourself

Let Y = RX . Which of the following is/are true:

Operator Notation: Check Yourself

Let Y = RX . Which of the following is/are true:

Operator Representation of a Cascaded System

Using operator notation: Y1 = (1 R) X Y2 = (1 R) Y1 Substituting for Y1 : Y2 = (1 R)(1 R) X

Equivalent operator expression: (1 R)(1 R) = 1 2R + R2

Delay Delay Delay

Equivalent operator expression: R(1 R) = R R2

Delay Delay Delay

Equivalent operator expression: (2 R) R(1 R) = (2 R)R (1 R)

How many of the following systems are equivalent?

How many of the following systems are equivalent?

Find y [n] given x[n] = [n]:

Find y [n] given x[n] = [n]:

Find y [n] given x[n] = [n]:

Find y [n] given x[n] = [n]:

Find y [n] given x[n] = [n]:

Find y [n] given x[n] = [n]:

Proof: Assume X2 = X1 : Y2 = (1 + R + R2 + R3 + ) X2 = (1 + R + R2 + R3 + ) X1 = (1 + R + R2 + R3 + ) (1 R) Y1 = ((1 + R + R2 + R3 + ) (R + R2 + R3 + )) Y1 = Y1 It follows that Y2 = Y1 .

Linear Dierence Equations with Constant Coecients