Simulations Assignments

Simulation 2 (Homework 4: Question 5)

Q5 (a) Genearate the data for LMS algorithm using the model H(z)= (z-0.8)(z+0.7)/{(z-0.9)(z+0.8)(z+65)}
To generate the data, you will drive the above system by a Gaussian white process with zero mean and variance preselected by you. Determine how long do you need to run the system, so that it has reached the steady state. Once in steady state, start collecting data. Collect, at least, 512 points.
(b) Get an estimate of signal energy for the above data, and using this estimate determine range for Mu. Select two values for Mu in this range.
(c) Run the LMS algorithm in predictive mode for the data you have generated and for the two choices of Mu.
(d) Do a validation test. You should use the following for the purpose of comparison
(i) Learning curve (i.e. Mean square error curve)
(ii) Convergent values of W(n)
(iii) Whiteness of error
Comment on which choice of Mu gives better results, and why.