6.

01 Final Exam
Name:

Spring 2011

Section:

Enter all answers in the boxes provided. Clearly written work will be graded for partial credit.
During the exam you may

refer to any printed materials use a calculator

You may not

use a computer, phone, or music player

For sta use: 1. 2. 3. 4. 5. 6. 7. 8. total: /8 /8 /6 /18 /11 /18 /15 /16 /100

6.01 Final Exam

Spring 2011

1 Alternative Excitations (8 points)

Find the relation between V1 and V2 that must hold so that I = 1A.

4 I V1

V2

Express the relation between V1 and V2 as an equation, and enter the equation in the box below.

equation:

6.01 Final Exam

Spring 2011

2 Lots of Feedback (8 points)

Consider the following system.

+ R +

Part a. Enter the dierence equation that relates the input sequence x[n] and output sequence y[n] for this system.

dierence equation:

Part b. Enter the pole(s) of the system in the box below. If there are multiple poles, separate them with commas. If there are no poles, enter none.

poles:

6.01 Final Exam

Spring 2011

Part c. Assume that you are given a feedback system whose performance is determined by a gain factor K. The following table species the two poles that result for six dierent values of K. The poles are rst described by their real and imaginary parts. The same poles are then also described by their magnitude and angles.

K 0.2 0.4 0.6 0.8 1 2

poles (Cartesion form) real imaginary

poles (polar form)

magnitude, angle

0.3 0.1j 0.6 0.2j 0.9 0.3j 1.2 0.4j 1.5 0.5j 3.0 1.0j

0.316, 0.632, 0.948, 1.264, 1.581, 3.162,

0.322 0.322 0.322 0.322 0.322 0.322

Which value of K could give rise to the following response?

K = 0.2 or 0.4 or 0.6 or 0.8 or 1 or 2 or none:

6.01 Final Exam

Spring 2011

3 OOP (6 points)
You are given the following (incomplete) denitions:
class MITstudent: GIR = [5.111, 7.011, 8.01, 8.02, 18.01, 18.02,
rest1, rest2, instlab,
hass1, hass2, hass3, hass4,
hass5, hass6, hass7, hass8]
DEPT = []
def __init__(self, AP = []):
self.taken = AP
def take(self, courses):
for course in courses:
self.taken.append(course)
def doneReq(self, req):
missing = []
for r in req:
if not r in self.taken:
print r + is missing
missing.append(r)
return len(missing) == 0
def done(self):
# your code here
class Course62(MITstudent): DEPT = [6.01, 6.02, 18.03, 6.041, 6.002, 6.003, 6.004, 6.005, 6.011, 6.013, 6.033, 6.115, 6.813, 6.336, 6.AUT, 6.AUP]

Note that the denition for each subclass of MITstudent, such as Course62, will dene a DEPT attribute specifying the departmental requirements. This code is intended to be used to track the courses taken by a student, and to test whether the student has satised all of the GIR and departmental requirements. A sample use is shown below:
Chris = Course62([18.01])
Chris.take([5.111, 18.02, 8.01, hass1])
Chris.take([18.03, 8.02, 6.01, hass2])
Chris.take([6.02, 6.041, 7.011, hass3])
Chris.take([6.002, 6.003, 6.013, hass4])
Chris.take([6.004, 6.005, 6.011, hass5])
Chris.take([6.013, 6.115, 6.336, hass6])
print Done? , Chris.done()
Chris.take([6.AUT, 6.033, 6.813, hass7])
Chris.take([6.AUP, random, random, hass8])
print Done? , Chris.done()

6.01 Final Exam

Spring 2011

3.1 Done
Write the done method of the MITstudent class so that it checks both the GIR and department requirements and returns a boolean (True if both requirements are complete; False otherwise).
def done(self):

3.2 Take
The previous denitions have one fundamental aw: they dont take into acount that 6.01 and 6.02 together satisfy the Institute Lab requirement instlab under the GIRs. We can address this aw by adding a take method to the Course62 class to handle this issue. Write the take method below; do not duplicate existing code.
def take(self, courses):

6.01 Final Exam

Spring 2011

4 Getting from A To B (18 points)

You are building a system to help people plan routes in a city, by a combination of walking and driving. Here is a class denition that characterizes the number of minutes to traverse a link via each of these modes.
class Link: def __init__(self, walkTime, driveTime):
self.walkTime = walkTime
self.driveTime = driveTime

Thus Link(20,10) species a link that takes 20 minutes to walk or 10 minutes to drive. If one of these options is not available (e.g., it is a footpath, or a freeway), then the time can be assigned a very large numerical value (called far). For example, Link(3,far) species a footpath that takes 3 minutes to walk and cannot be driven.

4.1 Personalized cost functions

Dierent people have dierent preferences about walking versus driving. One way to account for personal preferences is with costs, as follows. Let dm represent the driving multiplier. Then dene the cost of walking to be equal to the number of minutes in the walk, and the cost of driving to be dm times the number of minutes in the drive. Thus, if dm is 10, the cost of driving for 1 minute is the same as the cost of walking for 10 minutes. Similarly, if dm is 0.5, the cost of driving for 1 minute is the same as the cost of walking for 0.5 minutes. A cost function takes two inputs:
an object of type Link, and
a boolean indicating whether a car is available. It returns a numeric value, which is the cost of traversing that link. Note that if the car is not available, driving is not a valid option. Assume a person will choose to use whichever available transportation mode has the least cost. Write a procedure makeCostFunction(dm) that takes the driving multiplier for a person and returns a cost function for that person.
def makeCostFunction(dm):

6.01 Final Exam

Spring 2011

4.2 Going with the crowd

Now, assume a whole group of people want to go somewhere together. We would like to select a transportation modality that has the lowest cost averaged over the whole group. Write a procedure makeGroupCostFunction(costFunctions) that takes a list of cost functions (one for each of the people in the crowd) and returns a new cost function for the whole group. For full credit, use a list comprehension. You may assume a procedure average, which takes a list of numbers as input and returns the average.
def makeGroupCostFunction(costFunctions):

6.01 Final Exam

Spring 2011

4.3 Finding a path

Here is an example graph.

A
10, far 50, 20

B
50, 20 10, 2 1, far

C
far, 10

E
50, 20

D
30, 10 50, 20

G
Here it is, formalized in Python. It is like the maps we have used before, but instead of a single cost associated with each link, we have an instance of the class Link.
far = 1000000000 map1 = {A : [(B, (C, B : [(A, (D, (E, C : [(A, (D, (F, D : [(B, (C, (G, E : [(B, (G, F : [(C, (G, G : [(E, (D, (F, Link(10, far)), Link(50, 20))], Link(10, far)), Link(10, 2)), Link(50, 20))], Link(50, 20)), Link(1, far)), Link(far, 10))], Link(10, 2)), Link(1, far)), Link(30, 10))], Link(50, 20)), Link(50, 20))], Link(far, 10)), Link(50, 20))], Link(50, 20)), Link(30, 10)), Link(50, 20))]}

Assuming that the person always has the option to take a car, what is the best path for a person with dm = 0.1 from A to G? What is its cost? best path = associated cost =

6.01 Final Exam

Spring 2011

4.4 Domain model

Dene a class TransportationSearch, which is a subclass of sm.SM, suitable for use as input to ucSearch.smSearch. Its initializer should take a map and a cost function as input. You can just dene self.legalInputs to be [0, 1, 2, 3]. You do not need to specify the startState or the done test. Be sure to handle the case when the action is illegal (species a choice that does not exist) by remaining in the same state with a cost of 1. You can assume that the person always has the option of taking a car.
class TransportationSearch(sm.SM):

10

6.01 Final Exam

Spring 2011

4.5 Dude, wheres my car?

One problem with our model so far is that it assumes that a person can drive for a segment, walk for a segment, and then drive for another segment. Thats a problem, unless you can put your car in your pocket while you walk. So, we really need to stipulate that you can drive for any number of the initial segments of the path, but if you ever walk or take public transportation, you have to park your car, and cannot use it any further on this path. Briey describe a state and action space for this new version of the problem.

If the person has a car available initially and starts in location A, what is the initial state?

11

6.01 Final Exam

Spring 2011

5 Grid Search (11 points)

Consider a uniform grid that extends indenitely in all directions. The length between points in the vertical and horizontal directions is 1 meter. We start out at the grid point (state) marked S; the goal is far away and not shown in the gures.

In the grid on the left, in each state there are four actions that lead to four other states. In the one on the right there are eight actions per state. In all the questions, assume that the following two pruning rules apply. Pruning Rule 1: Do not consider a path that visits the same state twice. Pruning Rule 2: If multiple actions lead to the same state, consider just one of them. Note that a path of length 1 has exactly one action and a path of length 2 has exactly 2 actions.

5.1 Breadth First Search (BF)

Using BF in the 4-action grid, how many paths of length 1 are ever added to the agenda? Using BF in the 4-action grid, how many paths of length 2 are ever added to the agenda? Using BF in the 8-action grid, how many paths of length 1 are ever added to the agenda? Using BF in the 8-action grid, how many paths of length 2 are ever added to the agenda?

12

6.01 Final Exam

Spring 2011

5.2 Breadth First Search + Dynamic Programming (BF+DP)

Using BF+DP in the 4-action grid, how many paths of length 1 are ever added to the agenda? Using BF+DP in the 4-action grid, how many paths of length 2 are ever added to the agenda? Using BF+DP in the 8-action grid, how many paths of length 1 are ever added to the agenda? Using BF+DP in the 8-action grid, how many paths of length 2 are ever added to the agenda?

5.3 Uniform Cost Search (UC)

Assume that the cost of traversing a path is the Euclidean length of the path. Using UC (NO DP) in the 8-action grid,

What is the total cost of the second path expanded?

What is the total cost of the sixth path expanded?

What is the total cost of the tenth path expanded?

13

6.01 Final Exam

Spring 2011

5.4 Heuristics
The Manhattan distance between two states (x1 , y1 ) and (x2 , y2 ) on the grid is:

|x2 x1 | + |y2 y1 |
Assume that we are using A* search and we use this distance as an heuristic.

In the 4-action grid, are we guaranteed to nd the optimal path?

Yes/No?

Explain briey.

In the 8-action grid, are we guaranteed to nd the optimal path?

Yes/No?

Explain briey.

14

6.01 Final Exam

Spring 2011

6 Funky Elevator (18 points)

Youre riding in an elevator in a weird old building. It has three buttons, labeled up, down, and door-close; and it has a numeric oor indicator, which sometimes shows you a oor number and sometimes is turned o. There are 5 oors in the building. We are interested in estimating what oor the elevator is on, given the history of buttons you have pressed and numbers you have seen on the display. Your interaction with the elevator proceeds as follows: 1. 2. 3. 4. The elevator opens its doors (you may get on or o); The display indicates a oor number, or possibly fails to light up; You push a button; and The elevator moves to a new oor.

6.1 Nominal transitions

You have three possible actions: push the up button, push the down button, or push the door close button. The elevator was intended to work as follows: if you push the up button, then as a result, the elevator travels up one oor (but not above oor 4); if you push the down button, then as a result, the elevator travels down one oor (but not below oor 0); if you push the door-close button, then the elevator travels one oor in the direction specied by the last up or down button that was pushed. The states of the system are specied by the elevators oor and the direction that it is traveling. You are sure you enter the elevator on oor 2, but you think it is equally likely that the elevator is currently going up or going down. Indicate the initial belief state in the following table (states with probability of zero can be left blank). 0 up down 1 2 3 4

15

6.01 Final Exam

Spring 2011

6.2 Noisy transitions

In fact, this elevator is not so reliable. If you push the up button, then if it was already going up, it will go up one oor. If it was going down, then with probability 3/4, it will change directions and go up one oor, otherwise it will go down one oor. Similarly, if you push the down button, then if it was already going down, it will go down one oor; if it was already going up, then with probability 3/4, it will change directions and go down one oor, otherwise it will go up one oor. If you push the door-close button, it will move one oor in the direction it moved on the previous step. Starting from the initial belief state, what is the belief state of a state estimator after performing action door-close for 2 steps? 0 up down 1 2 3 4

Starting from the initial belief state, what is the belief state of a state estimator after performing action up and then performing action door-close? 0 up down 1 2 3 4

16

6.01 Final Exam

Spring 2011

Starting from the initial belief state, what is the belief state of a state estimator after performing action up for two steps? 0 up down 1 2 3 4

17

6.01 Final Exam

Spring 2011

6.3 Adding observations

When the elevator is on some oor f, there is a probability 0.2 that the display will be None, a probability 0.4 that the display will be f, probability 0.2 that it will display f + 1, and probability 0.2 that it will display f 1. This elevator is so crazy, that it will actually potentially display 1 or 5, so we dont need to worry about special handling for the case of the elevator being on oor 0 or 4.
Starting from the initial belief state, what is the belief state after the observation-action sequence:

[(None, door-close), (3, door-close)]

0 up down

Starting from the initial belief state, what is the belief state after the observation-action sequence:

[(3, door-close), (2, door-close)]

0 up down

18

6.01 Final Exam

Spring 2011

7 Position Controller (15 points)

The following gure shows a motor controller. A human can turn the left potentiometer (the input pot). Then the motor will turn the right potentiometer (the output pot) so that the shaft angle of the output pot tracks that of the input pot. VS VS

(1i )R i R

motor M+ M
19

(1o )R o R

The dependence of the pot resistances on shaft angle is given in terms of , which varies from 0 (most counterclockwise position) to 1 (most clockwise position). The resistance of the lower part of the pot is R and that of the upper part is (1 )R, where R = 1000. Notice that if i > o , then the voltage to the motor motor (VM+ VM ) is positive, and the motor turns clockwise (so as to increase o ) i.e., positive motor voltage clockwise rotation. Part a. Determine an expression for VM+ in terms of i , R, and VS . Enter your expression in the box below.

VM+ =

6.01 Final Exam

Spring 2011

Part b. The following circuit produces a voltage Vo that depends on the position of the input pot.

VS

R1 Vo R2 (1i )R i R

Determine an expression for the voltage Vo in terms of i , R, R1 , R2 , and VS . Enter your expression in the box below.

Vo =

20

6.01 Final Exam

Spring 2011

Part c. The following circuit produces a voltage Vo that depends on the positions of both pots.

VS R

(1i )R i R (1o )R o R Vo

Determine an expression for Vo in terms of i , o , R, and VS . Enter your expression in the box below.

Vo =

21

6.01 Final Exam

Spring 2011

Assume that we are provided with a circuit whose output is is possible to design a motor controller of the following form

i volts. We wish to determine if it o

+10V

motor i volts o M+ M

R4

R3 R2 R1 VC
so that the motorshaft angle (which is proportional to o ) will track the input pot angle (which is proportional to i ).

Part d. Assume that R1 = R3 = R4 = 1000 and VC = 0. Is it possible to choose R2 so that o

tracks i ?
Yes or No:
If yes, enter an acceptable value (a number) for R2 .

R2 =

If no, briey explain why.

22

6.01 Final Exam

Spring 2011

Part e. Assume that R1 = R3 = R4 = 1000 and VC = 5V. Is it possible to choose R2 so that o

tracks i ?
Yes or No:
If yes, enter an acceptable value (a number) for R2 .

R2 =

If no, briey explain why.

23

6.01 Final Exam

Spring 2011

8 Delay Removal (16 points)

Your company is interested in ensuring that the system shown below converges to its nal value as soon as possible. It has enough money to engineer one of the delays out of the system. Your job is to determine which delay, if removed, would result in the fastest convergence. Here is the original system. (It is also shown on the next page, for you to use as a worksheet).

1/2

R1 R2

R3

8.1 Analysis
Let H1 represent the system that results when R1 is a wire, R2 is a delay, and R3 is a delay. Let H2 represent the system that results when R2 is a wire, R1 is a delay, and R3 is a delay. Let H3 represent the system that results when R3 is a wire, R1 is a delay, and R2 is a delay. Fill in the following table with properties of H1 , H2 , and H3 . system function pole(s) converge? oscillate?

H1 =

H2 =

H3 =

8.2 Recommendation
Which delay should your company remove?
1 or 2 or 3 or none:
Briey explain.

24

6.01 Final Exam

Spring 2011

1/2

R1 R2

R3

25

6.01 Final Exam

Spring 2011

Worksheet (intentionally blank)

26

MIT OpenCourseWare http://ocw.mit.edu

6.01SC Introduction to Electrical Engineering and Computer Science

Spring 2011

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