Communication channel model

Goal
The goal of this experiment is to become familiar with the definition of channel model and the effect of the
channel model on the transmitted signal both in the time and the frequency domain.

Theory
Communicating data from one location to another requires some form of pathway or medium. These pathways,
called communication channels, use two types of media: cable (twisted-pair wire, cable, and fiber-optic cable) and
broadcast (microwave, satellite, radio, and infrared). Cable or wire line media use physical wires of cables to
transmit data and information. Twisted-pair wire and coaxial cables are made of copper, and fiber-optic cable is
made of glass.
The simplified block diagram of any communication system can be presented by Figure 1:

Figure 1. Block diagram of communication systems

As it is seen in Figure 1, in order to model the whole transmission environment, the transmitted signal first passes
through the channel and then, noise is added to the signal. In communication systems, noise is an error or undesired
random disturbance of a useful information signal in a communication channel. The noise is a summation of
unwanted or disturbing energy from natural and sometimes man-made sources. Noise is, however, typically
distinguished from interference, for example in the signal-to-noise ratio (SNR), signal-to-interference ratio (SIR)
and signal-to-noise plus interference ratio (SNIR) measures. Noise is also typically distinguished from distortion,
which is an unwanted systematic alteration of the signal waveform by the communication equipment, for example
in the signal-to-noise and distortion ratio (SINAD).

Various types of noise:

1. Thermal noise
Johnson–Nyquist noise (sometimes thermal, Johnson or Nyquist noise) is unavoidable, and generated by the
random thermal motion of charge carriers (usually electrons), inside an electrical conductor, which happens
regardless of any applied voltage.
Thermal noise is approximately white, meaning that its power spectral density is nearly equal throughout the
frequency spectrum. The amplitude of the signal has very nearly a Gaussian probability density function. A
communication system affected by thermal noise is often modeled as an additive white Gaussian noise (AWGN)
channel.

1
 The root mean square (RMS) voltage due to thermal noise 𝜗𝑛 , generated in a resistance R (ohms) over bandwidth
Δf (hertz), is given by

𝜗𝑛 = √4𝐾𝐵 𝑇𝑅∆𝑓 (1)

where 𝐾𝐵 is Boltzmann's constant (joules per kelvin) and T is the resistor's absolute temperature (kelvin).
As the amount of thermal noise generated depends upon the temperature of the circuit, very sensitive circuits
such as preamplifiers in radio telescopes are sometimes cooled in liquid nitrogen to reduce the noise level.
2. Shot noise
If electrons flow across a barrier, then they have discrete arrival times. Those discrete arrivals exhibit shot noise.
The output of a shot noise generator is easily set by the current. Typically, the barrier in a diode is used.
Shot noise in electronic devices results from unavoidable random statistical fluctuations of the electric current
when the charge carriers (such as electrons) traverse a gap. The current is a flow of discrete charges, and the
fluctuation in the arrivals of those charges creates shot noise. Shot noise is similar to the noise created by rain falling
on a tin roof. The flow of rain may be relatively constant, but the raindrops arrive discretely.
The root-mean-square value of the shot noise current 𝑖𝑛 is given by the Schottky formula

𝑖𝑛 = √2𝐼𝑞 ∆𝐵 (2)

where I is the DC current, q is the charge of an electron, and ΔB is the bandwidth in hertz.
The shot noise assumes independent arrivals. Vacuum tubes have shot noise because the electrons randomly
leave the cathode and arrive at the anode (plate). A tube may not exhibit the full shot noise effect: the presence of
a space charge tends to smooth out the arrival times (and thus reduce the randomness of the current).
Conductors and resistors typically do not exhibit shot noise because the electrons thermalize and move
diffusively within the material; the electrons do not have discrete arrival times. Shot noise has been demonstrated
in microscopic resistors when the size of the resistive element becomes shorter than the electron-phonon scattering
length.
3. Flicker noise
Flicker noise, also known as 1/𝑓 noise, is a signal or process with a frequency spectrum that falls off steadily
into the higher frequencies, with a pink spectrum. It occurs in almost all electronic devices, and results from a
variety of effects, though always related to a direct current.
4. Burst noise
Burst noise consists of sudden step-like transitions between two or more levels (non-Gaussian), as high as several
hundred microvolts, at random and unpredictable times. Each shift in offset voltage or current lasts for several
milliseconds, and the intervals between pulses tend to be in the audio range (less than 100 Hz), leading to the term
popcorn noise for the popping or crackling sounds it produces in audio circuits.

5. Transit-time noise

2
 If the time taken by the electrons from traveling from emitter to collector becomes comparable to the period of
the signal being amplified, that is, at frequencies above VHF and beyond, so-called transit-time effect takes place
and noise input admittance of the transistor increases. From the frequency at which this effect becomes significant
it goes on increasing with frequency and quickly dominates over other terms.
The noise that is considered to be added to the signal in communication systems is that thermal noise which is
modeled as additive white Gaussian noise (AWGN). Since this noise is completely random, it contains all the
frequencies as it is indicated in Figure 2.

Figure 2. AWGN noise contains all the frequencies

Channel modelling
Communication channels can be modeled as various types and each type has its own mathematical model both
in the time and frequency domains. In the following, some types of channel models are briefly discussed. In the
remainder of this instruction, x(t), y(t), h(t) and n define the transmitted signal, the received signal, channel impulse
response and noise all in the time domain. In addition, X(f), Y(f), H(f) and N define the transmitted signal, the
received signal, channel impulse response and noise all in the frequency domain.
1. Ideal channel
If signal is passed through an ideal channel, the received signal would be exactly the same as the transmitted
signal, 𝑦(𝑡) = 𝑥(𝑡) and 𝑌(𝑓) = 𝑋(𝑓). This scenario is impossible since the transmitted signal needs some time in
order to reach to the transmitter. In addition, in wireless communication conditions, the attenuation of the
transmitted signal in a loss free environment is proportional to the square of the distance between the transmitter
and the receiver. As a result what we mean by an ideal channel is the one that only delays and attenuates the
transmitted signal. The delayed and attenuated version of a signal is presented in Figure 3.

Figure 3. The bottom signal is the delayed and attenuated version of the upper signal after passing through the
medium

3
 The relation between the transmitter and the receiver in an ideal channel in the time domain is represented as:

𝑦(𝑡) = 𝑥(𝑡) ∗ ℎ(𝑡) = 𝑘𝑥(𝑡 − 𝜏) (3)

where 𝑘 and 𝜏 are attenuation and delay of the channel respectively. In the frequency domain we will have:

𝑌(𝑓) = 𝐻(𝑓). 𝑋(𝑓) = 𝑘𝑋(𝑓). 𝑒 𝑗2𝜋𝑓𝜏 (4)

Any other type of a channel is a non-ideal channel. In the following, some examples of non-ideal channels are
briefly reviewed.
2. Filtering channels
Any environment (like air, cox cables, fiber optics, etc.), would only be able to pass specific frequency band of
a signal. Based on the frequency band that the channel can pass, various types of channels would be defined as:
Low Pass, High Pass, Band Pass and Band Stop. These types are indicated in Figure 4:

Figure 4. Various types of filtering channels

The effect of low pass and high pass filters on a typical signal are illustrated in Figure 5.

Figure 5. Effect of the low pass and high pass filters on a typical signal

4
 As it is observed in Figure 5, low pass filter omits the fast variations of the signal and keeps the smooth
variations. In contrast, high pass filter, keeps the fast variation of the transmitted signal and omits the smooth
variation.
Filtering channels can be described in the frequency domain appropriately. Various types of filters can be
presented as:

𝐻𝑙𝑜𝑤−𝑃𝑎𝑠𝑠 (𝑓) = 1 − 𝑈(𝑓 − 𝑓𝑙𝑜𝑤−𝑝𝑎𝑠𝑠 )

𝐻𝐻𝑖𝑔ℎ−𝑃𝑎𝑠𝑠 (𝑓) = 𝑈(𝑓 − 𝑓𝐻𝑖𝑔ℎ−𝑝𝑎𝑠𝑠 )

(5)
𝐻𝐵𝑎𝑛𝑑−𝑃𝑎𝑠𝑠 (𝑓) = 𝑈(𝑓 − 𝑓𝑙𝑜𝑤−𝑝𝑎𝑠𝑠 ) − 𝑈(𝑓 − 𝑓𝐻𝑖𝑔ℎ−𝑝𝑎𝑠𝑠 )

𝐻𝐵𝑎𝑛𝑑−𝑆𝑡𝑜𝑝 (𝑓) = 1 − (𝑈(𝑓 − 𝑓𝑙𝑜𝑤−𝑝𝑎𝑠𝑠 ) − 𝑈(𝑓 − 𝑓𝐻𝑖𝑔ℎ−𝑝𝑎𝑠𝑠 ))

where 𝑈 defines the step function, 𝑓𝑙𝑜𝑤−𝑝𝑎𝑠𝑠 is the cutoff frequency for low pass filter and 𝑓𝐻𝑖𝑔ℎ−𝑝𝑎𝑠𝑠 is the start
frequency for high pass filter.
3. Multipath fading channels
In a crowded environment, the transmitted signal can reach to the receiver from various paths. Therefore, various
replicas of the transmitted signal arrives to the receiver and they interfere with each other at the receiver. An example
of the multipath environment is indicated in Figure 6.

Figure 6. The transmitted signal reach to the receiver from various paths

In Figure 7, it is indicated that how the summation of the shifted version of a signal causes interference at the
receiver.
Multipath fading channels can be described in the time domain appropriately. In the time domain, the relation
between the transmitted signal and the received signal is described as follows:

𝑦(𝑡) = 𝑥(𝑡) ∗ ℎ(𝑡) = 𝑘1 𝑥(𝑡 − 𝜏1 ) + 𝑘2 𝑥(𝑡 − 𝜏2 ) + ⋯ 𝑘𝐿 𝑥(𝑡 − 𝜏𝐿 ) (6)

where 𝑘1 , 𝑘2 , … 𝑘𝐿 are the attenuation of each path, 𝜏1 , 𝜏2 , … 𝜏𝐿 are the delay of each path and 𝐿 is the number of
paths.

5
 Figure 7. The summation of the shifted version of a signal causes interference

Equation (6) is defined by using the power delay profile (PDP) of ℎ(𝑡). PDP gives the intensity of a signal
received through a multipath channel as a function of time delay. The time delay is the difference in travel time
between multipath arrivals. The abscissa is in units of time and the ordinate is usually in decibels. It is easily
measured empirically and can be used to extract certain channel's parameters such as the delay spread. Put it in
simpler words, PDP defines the power and time distribution of the impulse response of a channel.
Any PDP is defined by the amplitude distribution and the delay distribution of the paths. Usually, time arrival
of the paths obey Poisson distribution. The most familiar power distributions are Rician and Rayleigh which are
used for line of sight (LOS) and non-line of sight conditions respectively. In LOS scenarios, the signal travels from
the transmitter to the receiver directly. As a result, it is more probable that it contains a higher amplitude in
comparison to the NLOS scenarios. Therefore, the signal’s amplitude is defined by Rician distribution which
considers more probability of having a higher amplitude. In contrast, in NLOS scenarios, in which the signal reaches
to the receiver after hitting to the obstacles and being attenuated, the power of the signal obeys Rayleigh distribution.
It means that the higher amplitude is less probable in NLOS scenarios.
It is important to notice that all the communication channels have filtering characteristics because all the
environments are only able to pass specific range of frequencies. As a result, a channel might be multipath fading
or not but it is always a filtering channel.
4. Nonlinear channels
All the channels that have been considered up to now, are linear channel. It means that the received signal is a
linear combination of the transmitted signal. Nonlinear channel model is also probable in communication systems.
One example of a nonlinear channel is the channel that multiply the transmitted signal to itself. In the time domain
we have:

𝑦(𝑡) = 𝑥(𝑡)2 (7)

and in the frequency domain we will have:

𝑌(𝑓) = 𝑋(𝑓) ∗ 𝑋(𝑓) (8)

The nonlinear channels are really undesirable since they make the reconstruction of the transmitted signal
challenging.