Simulations Assignments


  • Homework 2: Question 6 (Levinson Method)
  • Homework 4: Question 4
  • Homework 4: Question 5
  • Homework 5: Question 2 (Noise Cancellation)

    • Simulation 1 (Homework 2: Question 6)

    Question: Consider the sample values of a stationary process y(.) on the home page for the ASP course: in http://sharada.ee.iitb.ac.in/~ee608/, file name is levinson.dat
    (a) Determine the first 50 covariance lags r(n) 0<=n<=50
    > (b) Using Levinson's algorithm determine an approproate AR model. Comment on the goodness of this model. If there are any deficiencies or problem give reasons for them.

















































    Goodness of Prediction: Error covariance at the first iteration is 0.68741, which drops to 0.092287 in 5th itearation. But at the end of 100th itearation the error covariance remain at 0.0554. Hence, it can be concluded that algorithm converges very fast in first few iterations. After that there is no significant improvement.
    Also, from plot of predicted data, it is clear that it tracks actual data quite accurately.