The Cooper Union Department of Electrical Engineering ECE111 Signal Processing & Systems Analysis Final May 4, 2012

Similar documents
ECE2210 Final given: Fall 13

ECE137B Final Exam. There are 5 problems on this exam and you have 3 hours There are pages 1-19 in the exam: please make sure all are there.

EEL3135: Homework #4

Each problem is worth 25 points, and you may solve the problems in any order.

GEORGIA INSTITUTE OF TECHNOLOGY SCHOOL of ELECTRICAL & COMPUTER ENGINEERING FINAL EXAM. COURSE: ECE 3084A (Prof. Michaels)

Responses of Digital Filters Chapter Intended Learning Outcomes:

A system that is both linear and time-invariant is called linear time-invariant (LTI).

ECE2210 Final given: Spring 08

ECE4270 Fundamentals of DSP Lecture 20. Fixed-Point Arithmetic in FIR and IIR Filters (part I) Overview of Lecture. Overflow. FIR Digital Filter

ECE Circuit Theory. Final Examination. December 5, 2008

Massachusetts Institute of Technology

Problem Weight Score Total 100

UNIVERSITY OF CALIFORNIA College of Engineering Department of Electrical Engineering and Computer Sciences

Fourier Methods in Digital Signal Processing Final Exam ME 579, Spring 2015 NAME

GEORGIA INSTITUTE OF TECHNOLOGY SCHOOL of ELECTRICAL & COMPUTER ENGINEERING FINAL EXAM. COURSE: ECE 3084A (Prof. Michaels)

Digital Signal Processing I Final Exam Fall 2008 ECE Dec Cover Sheet

MAT 146. Semester Exam Part II 100 points (Part II: 50 points) Calculator Used Impact on Course Grade: approximately 30% Score

Massachusetts Institute of Technology

/ (2π) X(e jω ) dω. 4. An 8 point sequence is given by x(n) = {2,2,2,2,1,1,1,1}. Compute 8 point DFT of x(n) by

The Laplace Transform

Theory and Problems of Signals and Systems

EE538 Final Exam Fall 2007 Mon, Dec 10, 8-10 am RHPH 127 Dec. 10, Cover Sheet

Problem Value

NAME: ht () 1 2π. Hj0 ( ) dω Find the value of BW for the system having the following impulse response.

ECE 2210 Final given: Spring 15 p1

ECE 350 Signals and Systems Spring 2011 Final Exam - Solutions. Three 8 ½ x 11 sheets of notes, and a calculator are allowed during the exam.

EECE 2150 Circuits and Signals Final Exam Fall 2016 Dec 16

Roll No. :... Invigilator s Signature :.. CS/B.Tech (EE-N)/SEM-6/EC-611/ DIGITAL SIGNAL PROCESSING. Time Allotted : 3 Hours Full Marks : 70

Problem Value

Final Exam December 11, 2017

DON T PANIC! If you get stuck, take a deep breath and go on to the next question. Come back to the question you left if you have time at the end.

Controls Problems for Qualifying Exam - Spring 2014

ECE137B Final Exam. Wednesday 6/8/2016, 7:30-10:30PM.

ECE 3084 QUIZ 2 SCHOOL OF ELECTRICAL AND COMPUTER ENGINEERING GEORGIA INSTITUTE OF TECHNOLOGY APRIL 2, Name:

University of Illinois ECE 313: Final Exam Fall 2014

EE 40: Introduction to Microelectronic Circuits Spring 2008: Midterm 2

DSP Configurations. responded with: thus the system function for this filter would be

UNIVERSITY OF OSLO. Please make sure that your copy of the problem set is complete before you attempt to answer anything.

Laplace Transform Analysis of Signals and Systems

BENG 186B Winter 2012 Final

Probability and Statistics for Final Year Engineering Students

ECE538 Final Exam Fall 2017 Digital Signal Processing I 14 December Cover Sheet

Communications and Signal Processing Spring 2017 MSE Exam

University Question Paper Solution

Stability Condition in Terms of the Pole Locations

Continuous and Discrete Time Signals and Systems

Probability Space. J. McNames Portland State University ECE 538/638 Stochastic Signals Ver

SIDDHARTH GROUP OF INSTITUTIONS:: PUTTUR Siddharth Nagar, Narayanavanam Road QUESTION BANK (DESCRIPTIVE)

Problem Value Score No/Wrong Rec 3

ECGR4124 Digital Signal Processing Exam 2 Spring 2017

EE301 Signals and Systems Spring 2016 Exam 2 Thursday, Mar. 31, Cover Sheet

VALLIAMMAI ENGINEERING COLLEGE. SRM Nagar, Kattankulathur DEPARTMENT OF INFORMATION TECHNOLOGY. Academic Year


ELEN E4810: Digital Signal Processing Topic 11: Continuous Signals. 1. Sampling and Reconstruction 2. Quantization

EE 16B Final, December 13, Name: SID #:

DFT & Fast Fourier Transform PART-A. 7. Calculate the number of multiplications needed in the calculation of DFT and FFT with 64 point sequence.

FILTER DESIGN FOR SIGNAL PROCESSING USING MATLAB AND MATHEMATICAL

Machine Learning, Fall 2009: Midterm

55:041 Electronic Circuits The University of Iowa Fall Exam 2

EE301 Signals and Systems In-Class Exam Exam 3 Thursday, Apr. 20, Cover Sheet

Problem Weight Total 100

Question Bank. UNIT 1 Part-A

EECE 2150 Circuits and Signals Final Exam Fall 2016 Dec 9

Final Exam of ECE301, Prof. Wang s section 8 10am Tuesday, May 6, 2014, EE 129.

Final Exam December 20, 2011

Designing Information Devices and Systems II Spring 2016 Anant Sahai and Michel Maharbiz Midterm 2

EECE 2150 Circuits and Signals Final Exam Fall 2016 Dec 12

Chapter 4E - Combinations of Functions

EECE 301 Signals & Systems Prof. Mark Fowler

Prof. Dr.-Ing. Armin Dekorsy Department of Communications Engineering. Stochastic Processes and Linear Algebra Recap Slides

EE538 Final Exam Fall :20 pm -5:20 pm PHYS 223 Dec. 17, Cover Sheet

Problem Value Score No/Wrong Rec 3

Solution 7 August 2015 ECE301 Signals and Systems: Final Exam. Cover Sheet

Calculus I Sample Exam #01

Analysis of Finite Wordlength Effects

Module 4. Related web links and videos. 1. FT and ZT

ECE 650 Lecture #10 (was Part 1 & 2) D. van Alphen. D. van Alphen 1

ECE 438 Exam 2 Solutions, 11/08/2006.

ECE Branch GATE Paper The order of the differential equation + + = is (A) 1 (B) 2

ECGR4124 Digital Signal Processing Midterm Spring 2010

Exam 2 MAS 3105 Applied Linear Algebra, Spring 2018

EE482: Digital Signal Processing Applications

Basics on 2-D 2 D Random Signal

Stochastic Processes. M. Sami Fadali Professor of Electrical Engineering University of Nevada, Reno

for valid PSD. PART B (Answer all five units, 5 X 10 = 50 Marks) UNIT I

Massachusetts Institute of Technology

EEM 409. Random Signals. Problem Set-2: (Power Spectral Density, LTI Systems with Random Inputs) Problem 1: Problem 2:

DSP Design Lecture 2. Fredrik Edman.

MASSACHUSETTS INSTITUTE OF TECHNOLOGY Department of Electrical Engineering and Computer Science Discrete-Time Signal Processing Fall 2005

Lecture 2 OKAN UNIVERSITY FACULTY OF ENGINEERING AND ARCHITECTURE

Name: Student ID: Instructions:

Electronics II. Midterm II

ECE 202 Fall 2013 Final Exam

Discrete Simulation of Power Law Noise

Dynamic circuits: Frequency domain analysis

Review of Linear Time-Invariant Network Analysis

Homework Assignment 11

Roundoff Noise in Digital Feedback Control Systems

QUIZ ON CHAPTER 4 APPLICATIONS OF DERIVATIVES; MATH 150 FALL 2016 KUNIYUKI 105 POINTS TOTAL, BUT 100 POINTS

Parametric Signal Modeling and Linear Prediction Theory 1. Discrete-time Stochastic Processes

Transcription:

The Cooper Union Department of Electrical Engineering ECE111 Signal Processing & Systems Analysis Final May 4, 2012 Time: 3 hours. Close book, closed notes. No calculators. Part I: ANSWER ALL PARTS. WRITE YOUR ANSWERS IN THE SPACE PROVIDED HERE, NOT IN YOUR EXAM BOOKS. TOTAL: 30 points. Any errors will cause disproportionate penalty. IA: FORMULAS 1. Laplace transform. 2. z-transform. 3. Continuous-time convolution. 4. Discrete-time convolution. 5. CTFT. 1

6. Inverse CTFT. 7. DTFT. 8. Inverse DTFT. 9. DFT. 10. Inverse DFT. 2

IB: Basic Concepts 1. An analog transistor amplifier has simple poles at 2, 3 ± 4j and a double pole at 1; also zeros at 4 ± j and at +3. True or False (no justification): The system is stable. 2. A digital filter realized as a cascade of several direct form II transposed sections has poles at 1 2 e±jπ/3, ± 2j, and double poles at 1 3 3 e±j2π/3. It also has zeros at 2 3 e±jπ/4, 2, 3, and double zeros at 1 2 e±jπ/2. True or False (no justification): The system is stable. 3. BIBO stands for (do not define it; write what the acronym stands for): 4. The Laplace transform of 3δ (t) + 4e 3t u (t) is (no need to simplify): 5. The discrete-time unit step u [n] is defined as: 6. True or False (no justification needed): A transversal filter is always FIR (can not be IIR). 7. An LTI discrete-time system is FIR if its response has finite, and is IIR if its response has infinite. 3

IC: Digital Filter Structures Draw the filters on THIS PAGE in the space provided below. 1. Draw a transversal filter realization for: 2. Draw a direct-form II realization for: H (z) = 4 + 3z 1 + 2z 2 6z 3 H (z) = 4z2 + 2z + 1 2z 2 4z 3 4

Part II: Answer all parts. Points per problem indicated. Partial credit will be granted proportionately when warranted. 1. [8 pts.] Consider the following analog transfer function: s + 2 H (s) = K (s + 0.5) (s + 10) 2 First select K so that the DC gain is 20dB. Using that value for K: graph the Bode magnitude plot. Specifically, graph the straight-line asymptotes, determine corrections and sketch the actual curve. Be sure to label your diagram carefully. 2. [3 pts.] Given the following analog transfer function: H (s) = s (s + 30) (s + 10) 2 (s + 100) (s + 200) Specify the phase (in degrees) at DC and at. 3. [6 pts.] Given the following transfer function of an analog system: s + K s 2 + (4 + K) s + 9 where K is a variable gain parameter. It is real, but may be positive, negative, or zero. Assume K is restricted, as necessary, so that the system is stable. (a) Compute the natural frequency ω n and damping factor ζ, in terms of K. (b) Specify the condition on K for the system to be stable. (Remember, do not reject negative values unless you have to!) (c) For an underdamped system, specify the constraint on ζ and ALSO the corresponding constraint on K. Repeat for critically damped, and overdamped cases. 4. [3 pts.] Draw a sketch of the typical unit step responses for an underdamped system, and for an overdamped system. In particular, the rise time must be shown, correctly, to be less in one case than the other. Nevertheless, an engineer may select the system with the slower rise time if the preference is to have a system that behaves faster- BRIEFLY indicate what this means by pointing out features on the sketches. 5

5. [5 pts.] An oscillator circuit exhibits a spectral peak at 10M Hz. The spectrum is 3dB below peak at 8MHz and 12MHz. In the following, you can give explicit formulas for the answers, with numbers ready to be plugged into a calculator. (a) Specify the center frequency, bandwidth and Q. (b) Did you compute Q using an exact or approximate definition? DON T write out the other definition (the one you didn t use). (c) Do you have enough information to know if this is a second order circuit or not? What do you base your answer on? (d) Based on your answer to (c), for this case, do the exact and approximate definitions of Q give the exact same answer? 6. [10 pts.] Let H (s) be a 2 3 transfer function matrix of an analog system, with state-space realization {A, B, C, D}. The eigenvalues of A are 2 with multiplicity 3, and 1 ± j4 (each simple). (a) How many inputs, outputs and state variables are there? (b) Is the system internally stable? Is it externally stable? For each case, the answer is yes, no, or not enough information to be certain. No justification needed. (c) The matrix that relates the state at time t to the state at time 0, under condition of, is called the matrix (two words). In terms of A, this matrix is. The Laplace transform of this matrix is. (d) Referring to the same matrix as in part (c) above: in this case, based on the known information about A, write an explicit formula (in the time domain) for a typical entry, in terms of arbitrary constants. I intend for you to use modal analysis here; I don t mean set up the computation via matrix functional calculus. (e) Can H (s) have a pole at 4? Can H (s) NOT have a pole at 2? these would be an example of a hidden pole? Which of (f) Suppose we know that this is a minimal realization. Let {A, B, C, D } be any other realization (may or may not be minimal). What is known of the dimensions of A, B, C, D? (g) The H 23 (s) element of H (s) is the transfer function from the input to the output. 6

7. [6 pts.] Let H (z) be the transfer function matrix of a digital system, with state-space realization {A, B, C, D}. (a) Write the state space equations, with u [n] input vector, y [n] output vector and x [n] state vector. (b) Write the formula for H (z). (c) We can write x [n] =?x [0] (fill in the matrix) under the condition that. (d) Regarding part (c), x [n] 0 for all initial conditions x [0] if the of A are (specify a region of the complex plane). 8. [6 pts.] Let A be a 4 4 matrix. We want to determine a formula for A n, n 0, using matrix functional calculus. (a) We compute A n by first finding a polynomial r (λ) = r M λ M + + r 1 λ + r 0, and then making a particular substitution for λ. What should be M? (b) True or false (no justification): knowing the eigenvalues of A (with multiplicities), not the particular entries of A, is suffi cient to find the r k s. (c) Which is correct? The r k s will be functions of n, or the entity we substitute in for λ will depend on n. 9. [8 pts.] The complex-valued WSS random process x [n] satisfies the condition that: v [n] = x [n] 0.2x [n 1] + 0.3x [n 2] + 0.4v [n 1] + 0.1v [n 2] (1) results in white noise v [n] with σ 2 v = 3. time. Assume x and therefore v are 0-mean for all (a) Write an explicit formula for the PSD S x (ω) of x. (b) This filter, with x input and v output, is called the filter of x. The inverse is called the filter of x. (c) Actually, in parts (b), to be technically correct you are assuming the filter in (1) has a particular property, named (two words). This requires (only the poles / only the zeros / both poles and zeros) of (1) to be inside the unit circle. Don t actually check this! Just tell me what WOULD need to be checked. (d) v is called the signal of x. (e) Extra credit (10 pts for answering all parts): 1. You should be able to determine the values of E (x [n 1] v [n]) and E (x [n 2] v [n]) easily. What are they, and (briefly) why? (Clear answer in words is fine) 2. Use (1) and (i) to compute E (x [n] v [n]). 3. Verify that x [n] v [n] v [n]. 10. [3 pts.] Let x (t) be WSS with correlation r (τ). Assuming y (t) = x (t) exists, show that r yy (τ) = r xx (τ). Hint: Use the Wiener-Khinchin theorem and invoke a particular property of the Fourier transform you should know (no, don t write down the Fourier transform integral). 7

11. [4 pts.] In each case, specify all conditions that apply. If none apply, say so. No justification. (a) x (t) is Gaussian WSS and y (t) = x 2 (t). The output is (WSS/SSS/Gaussian). (b) x (t) is Gaussian WSS and y (t) = e t x (t). The output is (WSS/SSS/Gaussian). (c) x [n] is WSS and y [n] = Q [x [n]] where Q [ ] is a quantizer that rounds x to the nearest integer. Then y [n] is (WSS/SSS). (d) x [n] is SSS and y [n] = Q [x [n]] where Q [ ] is a quantizer that applies saturation overflow with a fixed level A: Q [α] = A if a A, and Q [α] = α if α < A. Then y [n] is (WSS/SSS). 12. [3 pts.] Write what the following abbreviations stand for (definitions not needed): (a) PSD (b) WSS (c) SSS [ ] [ ] X 2 13. [5 pts.] Consider the real random vector. The mean vector is µ = and Y 1 the correlation matrix is (partially) determined as: [ ] 8? R =? 5 (a) Specify the unknown entries that would GUARANTEE that X, Y are orthogonal. If it is not possible to force this just by specifying these values, say so. But no justification needed. (b) Specify the unknown entries that would GUARANTEE that X, Y are uncorrelated. As above, if you can t force this condition, say so. (c) Specify the unknown entries that would GUARANTEE that X, Y are independent. Same as above if you can t. (d) Suppose you know for certain that X, Y are jointly Gaussian. NOW repeat (a),(b),(c): DO ANY OF YOUR ANSWERS CHANGE? HOW? (e) Now suppose the unknown entries are GIVEN to you. Not necessarily any of the cases above, just generic values. Suppose you know X and Y are EACH Gaussian (that s all you know). Do you have enough information now to write down the pdf for 3X + 4Y? No justification needed. 8