Problem 12.1
a)
Decreasing the step size slow the Rate of convergence of the algorithm.
Mu=0.01
Mu=0.1
Increasing the step size will improve the convergence rate but would result in inaccurate results.
Due to sampling at incorrect times incorrect symbol v
Problem. 1
We can use other pulse shapes as well. For example, sinusoidal waves and sinc wave etc.
Sine shaped pulse.
This gives us smaller bandwidth. Hence, its possible to get narrow bandwidth pulse shapes.
Sinc function.
Bandwidth of sinc function.
Cos
Problem 12.3
Initial Time off set =-1
Eye opens near 1000 iterations. It opens at its widest between 1400-1450 iterations. Convergent value is
approximately 1.
Initial Time offset=-0.5, offset estimate=0.5.
Initial Time offset=-0.8, offset estimate=0.8.
I
Problem 11.5
% eyediag.m plot eye diagrams for pulse shape ps
N=1000; m=pam(N,6,12);
% random signal of length N
M=20; mup=zeros(1,N*M);
% oversampling factor of M
mup(1:M:N*M)=m;
% oversample by M
ps=hamming(M);
% hamming pulse of width M
x=filter(ps,1,m
Problem 5.6
Phi=0
Phi= -pi
Phi= -pi/2
Phi= -pi/3
Phi= -pi/6
Phi= pi/6
Phi=pi/3
Pi/2
Phi=pi
Depending on the phase the recovered messages spectrum may be entirely different from that of the
transmitted message. The problem gets severe when phase is +pi/2 o