27 Random Processes

271 Definition of a Random Process

Arandom process or stochastic process is an ensemble or
sample space. Such sample space
amsists of sample points whose outcome is random function of time.
Fig. 2.7.1 ilustrates the concept presented in above definition. (See Fig. 2.7.1 on next
page).
Here totality of x,(), x, (t) and x,(t) is called random process. Note that it is function of
time. x (t), x,(),..., x,(t) are the outcomes of points in sample space (ensemble). At t = t
the set of numbers x,(ty ), x,(t), ..., x,t )is a random variable.
27.2 Difference between Random ariable and Random Process
Following table illustrates the differences between the random variable and random
process.
Sr. No. Random Variable
Random Process
1 It is waveform.
it is a set of numbers.
2.
t need not be function of time. It is the function of time.
3. stationary or ergodic.
Random variables are not further classified. Random processes can be
4 can be
Only ensernble averages can be calculated. Ensemble as well as time averages
calculated.
random process
Table 2.7.1 Difference between random variable and it is random
When the value of the random process is taken at
particular time,
Variable. Thus random Process is made up from random variables.
 X(t) ix()
2

x(t)

t=t7
Fig. 2.7.1 Concept of random process
2.7.3 Mathematical Representation of Random Process
Let us assume that the interval- Tsts
the random process is random process is observed ov an

Xt, s), -T<tsTmathematically represented as,

Here 's
representsthe sample
S =,), sg = sample point in the And
X,),...S, =x() are the outcomes of att=
sample points.
Above set of
nte

processnumbers
The

the random on RHS number.

is
simplified andbasically
it is represents
simply
a random
written as X().
Digital S

Classification of of Random Processes

274
Depending upon the statistical characteristics and their
relationship with time, the
processes are classified as follows
nndom
1. Stationary and non-stationary random processes
2 Wide sense stationary or weakly stationary processes
3. Ergodicrandom processes
next sections.
These prOcesses are discussed in detail in

28 Stationary and Non-stationary Process

Processes
28.1 Definition of Stationary and Non-stationary
properties of a process do not change with time, it is called
When the statistical if the statistical properties are
said to be non stationary
IOTATY process. A process is generated from stable physical
process is
Tnction of time. Normally the stationary process is generated from
state. The non stationary
etomenon that is under steady
stable physical phernomenon.
 2.9.6 Wide
Sense
The prOcess mnay Stationary (or Weakly
..29.13
not be
stationary
hnctions are_ independent of shift of in strict
Stationary Process)
sense, still the mean and
stationary process. time origin. Such process is calledautocorrelation
wide sers
All processes in practice are non
stationary
end. This means statistics of such process are because every process has some start and
dependent on time. A truly stabionarv
process should start at t =-o and should not stop tillt= . Such type of proess is nt
possible prachcally. The process may appear stationary over a certain period of time. Then
ratoe
this will be wide sense stationary process.
 Distribution
.4 Gaussian or Normal cst
It is defined for Contut
Gaussian distribution is also called Normal Distribution. P
random variables. The PDF for a Gaussian random variable is given as,

Gaussian PDF: f(*) =

1 e-(x -m)²/2g2 ..25
o2
In the above equation
m = Mean of the random variable and
G' = Variance of the
Fig. 2.5.3 shows the plot of fy()
random variable. isobtai
for g = 0.25 and m = 1.23. with respect to x. The plot of this ngue along
X-axis. One feature of the Depending upon the values of m, the plot w
defines the spread of bell Gaussian PDE is that it looks like a bell stalue
shape from
mean Siep
 2.0

1.5
Area =0.19

1.0

Total area = 1.0

0.5

0.5 1.0 1.5 2.0

Fig. 2.5.3 Plot of Gaussian PDF

Properties of Gaussian PDF

Toperty 1 : The peak value occurs at x =m(i.e. mnean value). i.e.,

1 i.e. mean value ... (2.5.20)
at
fg(a) =
2.5.19.
This value is obtained by putting xX =m in equation
has even symmetry around mean value i.e.,
Gaussian PDF (2.5.21)
Pery 2: The plot of
- -= fr(m +o)
fr(m -o)
for all values of x below mean value
1/2
the PDF curve is
Prope
an3: The area under
and for all values of x above mean value.
i.e., (2.5.22)
1/2
1
PX <m) PX >m) = 2
impulse) function
aches to 8 (ie, And the