Correlation of DT signals
A mathematical operation that closely resembles convolution is correlation. In convolution the input
and impulse response are involved whereas in correlation two signal sequences are involved.
The correlation between the two signals is to measure the degree to which the two signals are similar
and thus to extract some information that depends to a large extent on the application.
Correlation of signals is often uncounted in radar, sonar, digital communications, geology and other
areas in science and engineering.
19 Lecture materials on "DiscreteTime Signals and Systems" By- Mohammed abdul kader, Assistant Professor, EEE, IIUC
Application of Correlation
Correlation in radar and active sonar applications
x(n) is the transmitted signal and y(n) is
the received signal.
If a target is present y(n) will be
Where 𝛼 is some attenuation factor representing
the signal loss involved in the round-trip
transmission of the signal x(n), D is the round trip
delay and w(n) represents additive noise that is
picked up by the antenna and any noise generated
by the electronic components and amplifier.
If there is no target,
y(n)=w(n)
Comparing two signal x(n) and y(n) radar detects whether a target is present or not and also
calculate the distance if target is present. In practice, the signal x(n-D) is heavily corrupted by the
additive noise to the point where a visual inspection of y(n) does not reveal the presence or
absence of the desired signal reflected from the target. Correlation provides us with a means for
extracting this important information from y(n).
20 Lecture materials on "DiscreteTime Signals and Systems" By- Mohammed abdul kader, Assistant Professor, EEE, IIUC
Application of Correlation
Correlation in radar and active sonar applications (Cont.)
21 Lecture materials on "DiscreteTime Signals and Systems" By- Mohammed abdul kader, Assistant Professor, EEE, IIUC
Application of Correlation (Cont.)
In digital Communication
Digital communication is another area where correlation is often used. In digital communications the
information to be transmitted from one point to another is usually converted to binary form, that is, a
sequence of zeros and ones, which are then transmitted to the intended receiver.
Signal sequence to transmit a logic 0: 𝑥0 (𝑛) for 0 ≤ 𝑛 ≤ 𝐿 − 1
Signal sequence to transmit a logic 1: 𝑥1 (𝑛) for 0 ≤ 𝑛 ≤ 𝐿 − 1
L represents the number of samples in each sequence.
The received signal can be represented as-
𝑦 𝑛 = 𝑥𝑖 𝑛 + 𝑤 𝑛 , 𝑖 = 0,1, 0≤𝑛≤𝐿−1
w(n) represents the additive noise.
After receiving y(n), the receiver compares the received signal y(n ) with both 𝑥0 (𝑛) and 𝑥1 (𝑛) to
determine which of the two signals better matches y(n). The comparison process is performed by means
of the correlation operation.
22 Lecture materials on "DiscreteTime Signals and Systems" By- Mohammed abdul kader, Assistant Professor, EEE, IIUC
Crosscorrelation and Autocorrelation
Crosscorrelation
Suppose that we have two real signal sequence x(n) and y(n) each of which has finite energy. The
crosscorrelation of x(n) and y(n) is a sequence 𝑟𝑥𝑦 (𝑙) , which is defined as
∞
𝑟𝑥𝑦 𝑙 = 𝑥 𝑛 𝑦 𝑛−𝑙 , 𝑙 = 0, ±1, ±2, … …
𝑛=−∞
Or, equivalently, as
∞
𝑟𝑥𝑦 𝑙 = 𝑥 𝑛+𝑙 𝑦 𝑛 , 𝑙 = 0, ±1, ±2, … …
𝑛=−∞
Autocorrelation
In special case where both signal sequences are same (i.e. y(n)=x(n)), we have the
autocorrelation of x(n), which is defined as the sequence
𝑟𝑥𝑥 𝑙 = 𝑥 𝑛 𝑥 𝑛−𝑙 , 𝑙 = 0, ±1, ±2, … …
𝑛=−∞
23
Lecture materials on "DiscreteTime Signals and Systems" By- Mohammed abdul kader, Assistant Professor, EEE, IIUC
Difference between Correlation and Convolution
In the computation of convolution, one of the sequence is folded, then shifted, then
multiplied by the other sequence to form the product sequence for that shift, and finally, the
values of the product sequence are summed.
Except for the folding operation, the computation of the crosscorrelation sequence involves
the same operation: shifting one of the sequence, multiplying the two sequence, and
summing over all values of the product sequence.
So, if we first fold a sequence y(n) to y(-n) and find the convolution between two sequences
x(n) and y(-n), it results crosscorrelation between x(n) and y(n).
𝑟𝑥𝑦 𝑙 = 𝑥 𝑙 ∗ 𝑦(−𝑙)
24
Lecture materials on "DiscreteTime Signals and Systems" By- Mohammed abdul kader, Assistant Professor, EEE, IIUC
Example of determining Correlation Sequence
Determine the crosscorrelation sequence 𝑟𝑥𝑦 (𝑙) of the sequences
𝑥(𝑛) = {2, −1,3,7,1,2, −3}
↑
𝑦(𝑛) = {1, −1,2, −2,4,1, −2, 5}
Solution: ∞
↑
𝑟𝑥𝑦 𝑙 = 𝑥 𝑛 𝑦 𝑛−𝑙 , 𝑙 = 0, ±1, ±2, … …
𝑛=−∞
n -4 -3 -2 -1 0 1 2 3 l
x(n) 2 -1 3 7 1 2 -3 0 l=0
𝑟𝑥𝑦 0 = 7
y(n) 1 -1 2 -2 4 1 -2 5 𝑟𝑥𝑦 −1 = 0
∑x(n)y(n) 2 1 6 -14 4 2 6 0 7 𝑟𝑥𝑦 −2 = 33
y(n+1) -1 2 -2 4 1 -2 5 0 l=-1
∑x(n)y(n+1) -2 -2 -6 28 1 -4 -15 0 0
y(n+2) 2 -2 4 1 -2 5 0 0 l=-2
∑x(n)y(n+2) 4 2 12 7 -2 10 0 0 33
25
Lecture materials on "DiscreteTime Signals and Systems" By- Mohammed abdul kader, Assistant Professor, EEE, IIUC
Example of determining Correlation Sequence (Cont.)
n -4 -3 -2 -1 0 1 2 l
x(n) 2 -1 3 7 1 2 -3
y(n+3) -2 4 1 -2 5 l=-3
𝑟𝑥𝑦 −3 = -14
∑x(n)y(n+3) -4 -4 3 -14 5 -14
𝑟𝑥𝑦 −4 = 36
y(n+4) 4 1 -2 5 0 l=-4 𝑟𝑥𝑦 −5 = 19
∑x(n)y(n+4) 8 -1 -6 35 0 36 𝑟𝑥𝑦 −6 = -9
𝑟𝑥𝑦 −7 = 10
y(n+5) 1 -2 5 0 l=-5
∑x(n)y(n+5) 2 2 15 19 For n<-7
y(n+6) -2 5 l=-6 𝑟𝑥𝑦 = 0
∑x(n)y(n+6) -4 -5 -9
y(n+7) 5 l=-7
∑x(n)y(n+7) 10 10
26
Lecture materials on "DiscreteTime Signals and Systems" By- Mohammed abdul kader, Assistant Professor, EEE, IIUC
Example of determining Correlation Sequence (Cont.)
n -3 -2 -1 0 1 2 l
x(n) -1 3 7 1 2 -3
y(n-1) 1 -1 2 -2 4 1 l=1
𝑟𝑥𝑦 1 = 13 ∑x(n)y(n-1) -1 -3 14 -2 8 -3 13
𝑟𝑥𝑦 2 = -18 y(n-2) 1 -1 2 -2 4 l=2
𝑟𝑥𝑦 3 = 16
∑x(n)y(n-2) 3 -7 2 -4 -12 -18
𝑟𝑥𝑦 4 = -7
𝑟𝑥𝑦 5 = 5 y(n-3) 1 -1 2 -2 l=3
𝑟𝑥𝑦 6 = -3 ∑x(n)y(n-3) 7 -1 4 6 16
For n>6
y(n-4) 1 -1 2 l=4
𝑟𝑥𝑦 = 0
∑x(n)y(n-4) 1 -2 -6 -7
y(n-5) 1 -1 l=5
The maximum similarity between two
signals x(n) and y(n) obtained when ∑x(n)y(n-5) 2 3 5
y(n) is delayed by 4 positions. y(n-6) 1 l=6
∑x(n)y(n-6) -3 -3
27
Lecture materials on "DiscreteTime Signals and Systems" By- Mohammed abdul kader, Assistant Professor, EEE, IIUC
