0% found this document useful (0 votes)
81 views5 pages

EE611 Solutions To Problem Set 1

This document provides solutions to problems involving orthonormal basis functions and signal representation. It determines orthonormal bases for given signals using Gram-Schmidt procedure and by inspection. It also verifies that the basis functions obtained from the two methods are related by a rotation matrix. The minimum number of bases required to represent the signals is determined. Finally, it represents signals in terms of orthonormal bases and provides a signal constellation diagram.
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
81 views5 pages

EE611 Solutions To Problem Set 1

This document provides solutions to problems involving orthonormal basis functions and signal representation. It determines orthonormal bases for given signals using Gram-Schmidt procedure and by inspection. It also verifies that the basis functions obtained from the two methods are related by a rotation matrix. The minimum number of bases required to represent the signals is determined. Finally, it represents signals in terms of orthonormal bases and provides a signal constellation diagram.
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 5

EE611 Solutions to Problem Set 1

1. The following four waveforms are given:

s0 (t) = rect(t) + rect(t − 2)


s1 (t) = rect(t − 1) + rect(t − 3)
s2 (t) = rect(t − 1) + rect(t − 2)
s3 (t) = rect(t − 1) − rect(t − 3)

(i) (a) Determination of orthonormal basis functions usimg Gram-Schmidt pro-


cedure starting with s0 (t) and going in sequence:
For i = 0, 1, 2, 3
i−1
X
gi (t) = si (t) − cik fk (t),
0

gi (t)
fi (t) = q ,
Egi
Z ∞
cik = si (t)fk (t)dt,
−∞
and Z
Egi = gi2 (t)dt.
Initialization: g0 (t) = s0 (t)
s0 (t)
f0 (t) = q
Es0
Es0 = 2
1
f0 (t) = √ [rect(t) + rect(t − 2)]
2
Determining the second basis function:

c10 = 0
g1 (t) = s1 (t) − c10 × f0 (t)
= rect(t − 1) + rect(t − 3)
⇒ Eg1 = 2

Therefore
g1 (t)
f1 (t) = q
Eg1
1
= √ [rect(t − 1) + rect(t − 3)]
2
Determining the third basis function:
1 1
c20 = √ , c21 = √
2 2

1
1
g2 (t) = [−rect(t) + rect(t − 1) + rect(t − 2) − rect(t − 3)]
2
Eg2 = 1
Therefore
1
f2 (t) = [−rect(t) + rect(t − 1) + rect(t − 2) − rect(t − 3)]
2
Determining the fourth basis function:

c30 = 0, c31 = 0, c32 = 1


1
g3 (t) = [rect(t) + rect(t − 1) − rect(t − 2) − rect(t − 3)]
2
Eg3 = 1

1
f3 (t) = [rect(t) + rect(t − 1) − rect(t − 2) − rect(t − 3)]
2

Using the basis functions {fi (t)}, the signals can be represented in vector
form as

s0 = [ 2 0 0 0]

s1 = [0 2 0 0]
1 1
s2 = [ √ √ 1 0]
2 2
s3 = [0 0 1 1].

(b) Choosing orthonormal basis functions by inspection:

b0 (t) = rect(t)
b1 (t) = rect(t − 1)
b2 (t) = rect(t − 2)
b3 (t) = rect(t − 3)

Using the basis functions {bi (t)}, the signals can be represented in vector
form as

s0 = [1 0 1 0]
s1 = [0 1 0 1]
s2 = [0 1 1 0]
s3 = [0 1 0 − 1].

2
(ii) In order to verify that one constellation can be obtained from the other by
a rotation, we express the basis functions {fi (t)} as a linear combination of
the basis functions {bi (t)}. Then, we show that the linear transformation C
is unitary, i.e., CT C = I.
1
f0 (t) = √ [b0 (t) + b2 (t)]
2
1
f1 (t) = √ [b1 (t) + b3 (t)]
2
1
f2 (t) = [−b0 (t) + b1 (t) + b2 (t) − b3 (t)]
2
1
f3 (t) = [b0 (t) + b1 (t) − b2 (t) − b3 (t)]
2
or equivalently,
√1 √1
 
0 0 b0 (t) f0 (t)
  
2 2
 0 √1 0 √1   b1 (t) f1 (t)
    
2 2 =
    
1 1 1   b (t) f2 (t)
− 21

 −2

2 2  2   
1
2
1
2
− 21 − 12 b3 (t) f3 (t)

i.e. Cb = f

where ith element of b is bi (t) and ith element of f is fi (t).

CT C = I.

(iii) The distance of each point from the origin and the relative distances between
the signal points remain the same in either case. (because of (ii)).

2. The minimum number of orthonormal basis functions required can be determined


using Gram-Schmidt procedure. The answer is 2.
Alternately, the minimum number of orthonormal basis functions can be determined
as follows:

• s0 (t) and s1 (t) are linearly independent. Therefore, at least 2 basis functions
are required.
1 1
• s2 (t) = s0 (t) − s1 (t) and s3 (t) = −s0 (t) − s1 (t), i.e., s2 (t) and s3 (t) can be
4 2
expressed as linear combinations of s0 (t) and s1 (t). Therefore, no more than
2 basis functions are necessary.

3. (a) The waveforms f1 (t), f2 (t), and f3 (t) are orthonormal because
(
1 i=j
Z ∞
fi (t)fj (t)dt = .
−∞ 0 i=
6 j

3
(b) If x(t) lies in the signal space spanned by f1 (t), f2 (t), and f3 (t), it can be
represented exactly as
3
X
x(t) = xi fi (t),
i=1

where Z ∞
xi = x(t)fi (t)dt.
−∞

If x(t) does not lie in the signal space spanned by f1 (t), f2 (t), and f3 (t), it can
be approximated as
3
X
x̂(t) = xi fi (t),
i=1

where Z ∞
xi = x(t)fi (t)dt
−∞

and Z ∞
[x(t) − x̂(t)]2 dt
−∞

is minimized.
In this case, we have x(t) = 2f1 (t) + f2 (t) − 3f3 (t).

4. One signal constellation representation of the two signals s1 (t) and s2 (t) is shown in
Figure 1. Gram-Schmidt procedure can be used to determine the basis functions.
√ √
A T A 3T
s2 = ( 2 , 2 )


s1 = (A T , 0)

Figure 1: Signal Constellation

5. The decision regions for the optimal receiver are shown in Figure 2.

4
y
y = 2x, Decision Boundary

(−1, 0.5)

(1, −0.5)

Figure 2: Decision Regions

You might also like