ECE 460: Communication and Information Theory Prof. B.-P. Paris Homework 5 Due: October 10, 2018
Reading
Madhow:
Chapter 3: sections 3.4 and 3.5. Note, the textbook in
section 3.5 covers synchronization in continuous time while we
treat it in discrete time in class. The latter is a little easier and
allows MATLAB experiments. Nevertheless, you need to read
section 3.5.
Vary the step size parameter μ to take the values 0.01, 0.05,
0.1, 0.25, 0.5, 0.75, 0.9, 0.95, 0.99. For each of these values,
measure the number of iterations n until |x[n]-2| < 10-5 for
the first time. Present your findings in a plot or table. Discuss
any observations that you deem interesting.
What did you observe for μ = 0.5? Can you explain your
observation based on the iterative update rule x[n + 1] =
x[n] ⋅ (1 - 2μ) + 4μ?
What do you observe when μ = 1? Can you explain your
observation based on the update rule?
What happens if μ < 0 or μ > 1? Can you explain your
observations from the update rule?
For this problem, disable frequency offset (set do_freq_offset =false) and disable the I part of the controller by setting eta =0.
Set the standard deviation of the additive noise to sigma=0.2;. Vary
the step size parameter μ to take on the values 0.01, 0.03, 0.1, 0.3, 1.
For each of these values of μ measure
the iteration number N0 such that the phase error for the
corrected signal r[n] is less than 0.01, i.e., |phase(r)| <
10-2, for the first time.
the variance of the phase error Var(phase(r)) =
∑n=N0+1N0+M|phase(r)|2, with M = 100.
Plot N0 versus the step size μ and plot varθ versus the step size μ.
Discuss your findings and make a well justified recommendation for
how the step size μ should be selected.
Repeat the experiment for two other noise levels (e.g.,
sigma=0.05; and sigma=0.8;. Would you change your choice of μ
in either of these cases?
Repeat the preceding problem with frequency offset and integral
control: set do_freq_offset = true and eta = mu/100.