Signals
By
R.N. Akol
Signals
• Qn -What is a signal?
– Quantity that varies with time and generally carries
information e.g. Voltage
– Signal is synonymous with a system
• A system extracts, modifies, and transforms the
information
• Hence the study of transforms
• A signal can be further classified in terms of its
dimension.
– We shall only look at signals with one dimension
Signal Classifications
• Periodic or non periodic
– Periodic Signals repeat after certain amount of time
• f(t+T)=f(t) where T is the period or 1/T is the fundamental
frequency and
• e.g. Sinusoidal
• The entire signal can be constructed from its period
• If two periodic signals follow the super position theorem i.e.
addition of two signals is also periodic
• The area under curve
– Non periodic (aperiodic) signals
• dfn does not repeat itself after a specific time interval.
Signals cont...
• Even or Odd signals
– Even signals are symmetric about the y-axis
For example, a cosine function
• f(t)=f(-t)
– Odd signal are symmetric about the origin
• f(t)=-f(-t)
• Continuous and discrete time signals
– Continuous signals are signal that are define/specified for all the time
• e.g temperature, light, etc
– Discrete signal are define at discrete time intervals
• e.g traffic lights, heart beat, clock etc
• Note : discrete signals can be derived from continuous signal through
analog to digital conversion
• Analog and Digital signal
– Analog signal is a continuous signal that takes on values in a given
range e.g. Sound wave, music, pictures
– Digital signal is a discrete signal that takes on distinct values (a finite
set of possible wave forms) e.g binary signal, quaternary, etc.
Signal cont...
• Real and complex signals
– Real signal is one with its values that are real numbers
– Complex signal takes on values in a set of complex
numbers
• A complex signal can be represented by two real signals i.e. real
and imaginary part or amplitude and phase values
– Most communication signals are complex in nature
conveying the amplitude and phase in the frequency
domain
• Deterministic and non-deterministic/ random signals
– Deterministic signals are signals whose values are
completely defined at any given time
• Deterministic signals can be modelled by a given function at any
given time
– Random signals........
• Most communication signals are random in nature
Other types of signals
energy signal- whose energy is finite and power is zero e.g. aperiodic signals
• Energy and power signals power signal- whose power is finite and energy is infinite e.g. periodic signals.
• Time or bandwidth limited signals time limited- its non zero for only a finite length of time.
band limited- its frequency components are zero above a
certain frequency. Its represented by a finite number of
• Casual and non casual signals
harmonics.
causal-zero for all negative time
non- causal- zero for all positive time.
• Baseband and bandpass signals
baseband- original message signal intended to be transmitted
bandpass- modulated/filtered signal which will be converted back to the base band signal
Random Signals
• Random signals take on random values at any given
time
– Rapidly fluctuate
– They evolve with time, distance or an given parameter
– Cannot be defined by a mathematical equation
– Characterized by probability and statistics
• Qn. Name some of the common probability functions
binomial
• Many communication signals are random normal
– This is mainly because deterministic signal do not carry
information
– Any information carrying signal must exhibit some level of
uncertainty
cont...
• A collection of signals is called a random process
• Each signal in the collection is referred to as a
sample space
• Probability models is an experiment with possible
outcomes chosen from a set of sample space
– The sum of all possible outcomes is equal to one
– Give examples.... tossing a coin or a dice
– The sample space is the probability measure of the
assemble
Useful statistical averages
• Mean, median & mode
• Average power
• Variance
• autocorrelation
Statistical averages
• Mean of the random process X(t) is the
expectation of the random variable
µx (t ) = E[ X (t )] = X (t )
– Find the mean of a continuous and discrete
random signal...... ???
• Power of a random signal is defined as
Px = E[ X (t )]
2
– Find the power of a continuous and discrete
random signal ......... ???
Statistical Average
• Variance
s x2 = E[ X (t ) - µ x ]2
– Find the variance of a discrete & continuous
random signal ???
• Auto correlation function is the product of
two random variables X(t1) & X(t2) observed
at different time intervals
Rx (t1 , t2 ) = E[ X (t1 ) X (t2 )]
Auto Correlation function(fn)
• For stationary random signals Rx (t2 - t1 ) = Rx (t )
• Write the auto correlation fn for a discrete &
continuous signals ???
• Properties
– The mean square value of a random process can be
obtained from the Rx (t ) by replacing t = 0 , Rx (0) = E[ X 2 (t )]
– The R (t ) is an even function
x
– The Rx (t ) has a maximum magnitude at t = 0
Statistical average
• Qn. What is the significance of:
– Mean if the expected value of the signal amplitude does not change with time, its a stationary process.
– Average power
– Variance measure of how the signal varies over time
• Qn. What are the characteristics of the
autocorrelation? ???
• Qn. What is the cross correlation?
???
Stationary and non stationary signal
• Random signals can further be classified as
stationary or non stationary signals.
– Stationary random signal
• The expected values of the signal does not change over time (
mean, average power)
– Non stationary random signal......
expected values change over time.
Transforms
• Fourier
• Hilbert
• Laplace
• Z transform
• Convolution
• Derivatives/integrals
Fourier transform/series
• Used to evaluate the frequency content of an
signal
• Frequency domain representation
– Makes large signal processes much faster
• All signals can be represented as a sum of
sinusoids (square & Triangular signals)
– Sinusoids are linear in the frequency domain
• Deals with signals in their steady state
– Periodic
Fourier cont...
• Representation Fourier series
¥ ¥
x(t ) = a 0 + å a n cos(wn t ) + å bn sin(wn t )
n=1 n=1
– Where the coefficients are given by
T
1
a 0 = ò x(t ) dt ,
T 0
T
2
an =
T ò x(t ) cos(w t )dt ,
0
n
T
2
bn =
T ò x(t ) sin(w t )dt
0
n
cont.....
• w = 2p f = 2p n fundamental frequency
n n
T
• n is the nth harmonic
• Fourier series defines the amplitude of the
harmonics in a periodic wave.
• Frequency spectrum of the wave (drawing both amplitude and phase spectra)
Fourier cont....
• Complex exponential series
¥
x(t ) = åx
n =-¥
n
exp(2p n t/ T),
T
1
xn = ò x(t ) exp(2p n t/ T)
T 0
Fourier cont....
• Fourier transform of continuous function
¥
F{x(t )} = ò x(t ) exp(- j2p n t/ T)dt
-¥
• Properties of the Fourier transform
Fourier Transform
• Name some properties of the Fourier
Transform
• Fourier transform pairs
Other functions in signal analysis
• Power spectral density
– Properties of the power spectral density
• Cross correlation
– Properties of the cross correlation
• Probability distribution function (PDF)
