Carnegie Mellon University

Department of Electrical and Computer Engineering

18-751

APPLIED STOCHASTIC PROCESSES

COURSE SYLLABUS FOR FALL 1999

WEB SITE

http://www.ece.cmu.edu/~ee751

This is where you will find most of the information about this course, including handouts, homework assignments, homework solutions, etc.

Lectures

MW 10:30AM-12:20PM           HH B103

Instructor

Prof. Tsuhan Chen          tsuhan@cmu.edu                x8-7536            Porter Hall B16

                                       Office Hours: MW 12:20-1:30PM or by email appointment

Teaching assistant (T.I.)

Deepak Turaga              dturaga@andrew.cmu.edu      x8-2528            Porter Hall B21

                                      Office Hours: TR 2:00-3:00PM or by email appointment

Course secretary

Carol Patterson             carol@ece.cmu.edu                x8-7286            Porter Hall B15

Course description

In this course, we introduce random processes and their applications.  Throughout the course, we mainly take a discrete-time point of view, and discuss the continuous-time case when necessary.  We first introduce the basic concepts of random variables, random vectors, stochastic processes, and random fields.  We then introduce common random processes including the white noise, Gaussian processes, Markov processes, Poisson processes, and Markov random fields.  We address moment analysis (including Karhunen-Loeve transform), the frequency-domain description, and linear systems applied to stochastic processes.  We also present elements of estimation theory and optimal filtering including linear prediction, Wiener and Kalman filtering.  Advanced topics such as linear models and spectrum estimation are discussed.  Applications of stochastic processes are studied and reviewed through student seminars.

Prerequisites

Probability Theory (such as 36-217) and Signals and Systems (such as 18-396) or equivalents.  Prior knowledge in discrete-time signal processing will be helpful.

Text

n         Discrete Random Signals and Statistical Signal Processing, Charles W. Therrien, Prentice Hall

References

n         Probability, random Variables, and Stochastic Processes , 3rd Edition, Athanasios Papoulis, McGraw-Hill

n         Probability, Random Variables, and Random Signal Principles, 3rd Edition, Peyton Z. Peebles, Jr., McGraw-Hill

n         Probability and Random Processes: Using MATLAB with Applications to Continuous and Discrete Time Systems, Donald G. Childers, McGraw-Hill/Irwin

n         Probability and Random Processes for Electrical Engineering, 2nd Edition, Alberto Leon-Garcia, Addison Wesley

HOMEWORK

Problem sets will be assigned every week on Wednesday, and will be due in class, before 10:30AM, on the succeeding Wednesday.  Expect about 10 problem sets.  A component of the homework will be computer assignments.  The homework policy is as follows:

n         You can discuss the homework problems with any number of students currently taking the course, the TA, and the instructor.  However, solutions and solution-techniques should not be exchanged.  You should make sure that you understand the solution you turn in, and should of course write up every word of the solution by yourself.  For problems which have an answer (Problems such as “Find the output” or “Plot the result”), it is OK to compare your final answer with others currently enrolled in the course.  But you should fix up any error by your own effort.  If these sentences are still vague, just tell yourself “I shall not take undue advantage of any other student” and this should answer other policy-related questions you have in your mind.

n         During the entire semester, do not look at the solutions to any homework or exams of previous years.

n         In order to work out the homework assignments, you should not look at any sources such as books, solutions manuals, papers, and other articles that are not mentioned or handed out in the class.  (The only exceptions are mathematical tables, and standard texts which you have used in your past career, while taking courses at CMU or elsewhere.)  Even though this might sound counter-educational, it is not.  We believe that it is much more educational to try out a proof or reasoning by yourself, rather than just grab a journal and read it up.  In any case, you get the leisure to work on such tutorial exercises by yourself only in a University!

n         All the answers you give must be fully justified.  However, results proved in the class, in class-given handouts, in past assignments, or in text can be used without proof, provided you specifically cite the source.

n         Sometimes you will find that a particular homework problem is apparently unrelated to the lectures.  This is intentional, the aim being to get acquainted with additional material that cannot be covered in class.

n         No late homework.

Student Seminars

Students will form groups (2~3 students per group) to study emerging applications of stochastic processes.  Each group will submit a written report and present a 10~15 min seminar in class (with a rehearsal with the instructor in advance).  These seminars will be scheduled in the second half of the semester.

Exams

There will be two exams, one mid-term and one final exam.  Exams will be conducted in class and are closed-book exams.

Grading

Homework:            40%

Seminars:               20%

Midterm exam:       20%

Final exam:             20%

THE FILE CABINET

The file cabinet under the exit sign near Porter Hall B18 is where you will find your graded papers in case you forget to pick them up in class.