Scribd
Upload
43 views

Com Solved Proble4ms I

1. The random process X(t) is nonstationary because the probability density functions of samples taken at different times t1 and t2 are not equal. 2. The random process X(t) = Acos(2πfct) is nonstationary because the probability density functions of samples taken at times ti and ti+τ are not equal, even though the frequency fc is constant. 3. The cross-correlation between the quadrature modulated processes X1(t) and X2(t) is zero at τ = 0 because the phase Θ is uniformly distributed, making the processes orthogonal (unrelated) to each other.

Uploaded by

NatyB
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
43 views

Com Solved Proble4ms I

1. The random process X(t) is nonstationary because the probability density functions of samples taken at different times t1 and t2 are not equal. 2. The random process X(t) = Acos(2πfct) is nonstationary because the probability density functions of samples taken at times ti and ti+τ are not equal, even though the frequency fc is constant. 3. The cross-correlation between the quadrature modulated processes X1(t) and X2(t) is zero at τ = 0 because the phase Θ is uniformly distributed, making the processes orthogonal (unrelated) to each other.

Uploaded by

NatyB
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

Solved Problems (Chapter I)

1. Consider a random process 𝑋(𝑡) is defined by


𝑋(𝑡) = sin⁡(2𝜋𝑓𝑐 𝑡)
In which the frequency⁡𝑓𝑐 is a random variable uniformly distributed over the interval [0,
W]. Show that 𝑋(𝑡) is nonstationary. Hint: Examine specific sample functions of the
random process 𝑋(𝑡) for the frequency 𝑓 = 𝑊⁄4 , 𝑊⁄2 , 𝑊, 𝑠𝑎𝑦.
Solution:
As an illustration, three particular sample functions of the random process⁡𝑋(𝑡),
corresponding to⁡𝑓 = 𝑊⁄4 , 𝑊⁄2 , 𝑊, are plotted.

To show that 𝑋(𝑡) is nonstationary, we need only observe that every waveform illustrated
1 1
above is zero at t=0, positive for 0 < 𝑡 < 2𝑊 , and negative for − 2𝑊 < 𝑡 < 0. Thus, the
probability density function of the random variable 𝑋(𝑡1 ) obtained by sampling 𝑋(𝑡) at
1
𝑡 = 4𝑊 is identically zero for negative argument, whereas probability density function of
−1
the random variable 𝑋(𝑡2 ) obtained by sampling 𝑋(𝑡) at 𝑡 = 4𝑊 is nonzero only for
negative arguments. Clearly, therefore,
𝑓𝑥(𝑡1 ) (𝑥1 ) ≠ 𝑓𝑥(𝑡2 ) (𝑥2 ),⁡⁡⁡⁡⁡⁡⁡⁡ and the random process is nonstationary.

1
2. Consider the sinusoidal process
𝑋(𝑡) = 𝐴𝑐𝑜𝑠(2𝜋𝑓𝑐 𝑡)⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡
Where the frequency 𝑓𝑐 is constant and the amplitude A is uniformly distributed:
1,⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡0 ≤ 𝑎 ≤ 1
𝑓𝐴 (𝑎) = {
0,⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡𝑂𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒⁡
Determine whether or not this process is strictly stationary.
Solution:
𝑋(𝑡) = 𝐴𝑐𝑜𝑠(2𝜋𝑓𝑐 𝑡)⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡
Therefore,
𝑋𝑖 = 𝐴𝑐𝑜𝑠(2𝜋𝑓𝑐 𝑡𝑖 )⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡
Since the amplitude A is uniformly distributed, we may write,
1
,⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡0 ≤ 𝑥1 ≤ 𝑐𝑜𝑠(2𝜋𝑓𝑐 𝑡𝑖 )
𝑓𝑥𝑖 (𝑥1 ) = {𝑐𝑜𝑠(2𝜋𝑓𝑐 𝑡𝑖 )
0,⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡𝑂𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒⁡

Similarly, we may write


𝑋𝑖+𝜏 = 𝐴𝑐𝑜𝑠[2𝜋𝑓𝑐 (𝑡𝑖 + 𝜏)]⁡⁡⁡⁡⁡⁡⁡
And
1
,⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡0 ≤ 𝑥2 ≤ 𝑐𝑜𝑠[2𝜋𝑓𝑐 (𝑡𝑖 + 𝜏)]
𝑓𝑥𝑖+𝜏 (𝑥2 ) = {𝑐𝑜𝑠[2𝜋𝑓𝑐 (𝑡𝑖 + 𝜏)]
0,⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡𝑂𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒⁡
We thus see that 𝑓𝑡𝑖 (𝑥1 ) ≠ 𝑓𝑡𝑖+𝜏 (𝑥2 ),⁡⁡and the random process 𝑋(𝑡) is nonstationary.
3. Consider a pair of quadrature modulated processes 𝑋1 (𝑡) and 𝑋2 (𝑡) that are related to a
stationary process as
𝑋1 (𝑡) = 𝑋(𝑡)⁡𝑐𝑜𝑠(2𝜋𝑓𝑐 𝑡 + 𝛩)⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡
𝑋2 (𝑡) = 𝑋(𝑡)⁡𝑠𝑖𝑛(2𝜋𝑓𝑐 𝑡 + 𝛩)⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡

2
Where 𝑓𝑐 is a carrier frequency, and the random variable 𝛩 is uniformly distributed over
the interval⁡[0, 2𝜋]. Moreover, 𝛩 is independent of 𝑋(𝑡). Find the cross-correlation
between 𝑋1 (𝑡) and⁡𝑋2 (𝑡).
Solution:

Where, in the last line, we have made use of the uniform distribution of the random variable
𝛩 representing the phase.
Note that, at 𝜏 = 0, the⁡factor ⁡sin(2𝜋𝑓𝑐 𝜏) is zero and therefore,

This shows that the random variables obtained by simultaneously observing a quadrature-
modulated processes 𝑋1 (𝑡) and⁡𝑋2 (𝑡) at some fixed value of time t are orthogonal
(unrelated) to each other.
4. Consider the random process 𝑋(𝑡) = 𝐴⁡𝑐𝑜𝑠(2𝜋𝑓𝑐 𝑡 + 𝛩), where the random variable 𝛩 is
uniformly distributed over the interval⁡[−𝜋, 𝜋]. The autocorrelation function of this random
process is given by the equation below (How?):

Find the power spectral density of the process.


Solution:
Let 𝛿(𝑓) denote the delta function at⁡𝑓 = 0. Taking the Fourier transform of both sides of
the relation defining⁡𝑅𝑥 (𝜏), we find that the power spectral density of the sinusoidal
process 𝑋(𝑡) is

Which consists of a pair of delta functions weighted by the factor 𝐴2 ⁄4 and located at⁡±𝑓𝑐 ,
as illustrated in the following figure. We note that the total area under the delta function is
one. Hence the total area under 𝑆𝑥 (𝑓) is equal to⁡𝐴2 ⁄2, as expected. (Property 1 of
Autocorrelation functions and Property 2 of PSD.)

3
5. Figure below shows the sample function x(t) of a random process X(t) consisting of a
random sequence of binary symbols 1 and 0. The following assumptions are made:
1. The symbols 1 and 0 are represented by pulses of amplitude +A and –A,
respectively and duration T seconds.
2. The pulses are not synchronized, so that the starting time⁡𝑡𝑑 of the first complete
pulse for positive time is equally likely to lie anywhere between zero and T seconds.
That is, 𝑡𝑑 ⁡is⁡a⁡sample⁡value⁡of⁡a⁡uniformly⁡distributed⁡random⁡variable⁡𝑇𝑑 ,
with its probability distribution function defined by
1
𝑓𝑇𝑑 (𝑡𝑑 ) = { 𝑇 ⁡,⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡0 ≤ 𝑡𝑑 ≤ 𝑇
0,⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡𝑒𝑙𝑠𝑒𝑤ℎ𝑒𝑟𝑒

3. Symbols 1 and 0 are equally likely, and thus the presence of a 1 or 0 in any one
interval is independent of all other intervals.
Find the mean and autocorrelation function of the process.

Solution:
Since the amplitude levels +A and –A occur with equal probability, it follows immediately that
E[X(t)]=0 for all t, and the mean of the process is therefore zero.

4
Let 𝑋(𝑡𝑘 ) and 𝑋(𝑡𝑖 ) be random variables obtained by observing the random process 𝑋(𝑡) at
times⁡𝑡𝑘 ⁡𝑎𝑛𝑑⁡𝑡𝑖 , respectively.
Consider first the case when⁡|𝑡𝑘 − 𝑡𝑖 | > 𝑇. Under this condition the random variables 𝑋(𝑡𝑘 ) and
𝑋(𝑡𝑖 ) occur in different pulse intervals and are therefore independent. We have thus,
⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡𝐸[𝑋(𝑡𝑘 )𝑋(𝑡𝑖 )] = 𝐸[𝑋(𝑡𝑘 )]𝐸[𝑋(𝑡𝑖 )] = 0, |𝑡𝑘 − 𝑡𝑖 | > 𝑇

5
6. For the above problem (6), determine its power spectral density.
Solution:

You might also like