J. Opt. Commun.

2020; aop

Sanmukh Kaur*

Performance analysis of FSO link under the effect

of fog in Delhi region, India
https://doi.org/10.1515/joc_2020-0151 raise problems of interference and hence usability of fre-
Received June 28, 2020; accepted August 17, 2020; published online quencies thus affecting the system capacity. RF spectrum is
October 16, 2020 becoming more congested and thus expensive to acquire as
a result of progressively increasing applications of RF
Abstract: Fog attenuation causes more loss than other
communication. Several efforts are being put by industry
weather conditions in a free space optics (FSO) link thus
and research for stretching the capability of existing wire-
limiting the visibility distance. This work presents a detailed
less technologies by alleviating interference and for devel-
survey on the attenuation of the transmitted signal as a
oping the new technologies capable of fulfilling the
result of variation in visibility range caused by different fog
emerging needs [3–5].
conditions of the Delhi, Safdarjung region. Kim and Kruse
Free space optical wireless communication (FSOWC)
models have been used to calculate attenuation as a result of
has been widely investigated over the past few years as an
fog conditions for three specific months (January, February
alternative to radio frequency technology. This technology
and December) for seven consecutive years starting from
is similar to optical fiber as input data or information signal
2013 to 2019. Received signal quality has been analyzed as a
is used for modulating a laser light beam in the transmitter.
function of transmitted power, data rate, transmission range
The modulated light beam then propagates in a wireless
and operation wavelength. Descriptive statistical analysis of
manner from transmitter to the receiver [6, 7]. The recent
real time observed visibility data allows for the estimation of
increasing interest in this type of communication stems
specific optical attenuation and enables in determining the
from the fact that it combines the higher bandwidth feature
link availability of the region for the complete year.
of fiber optics communication with the flexibility offered by
Keywords: atmospheric fog; attenuation; data rate; FSO; wireless communication technologies. Free space optics
Q-factor; transmission range; visibility. (FSO) transmission is extremely secure as a result of narrow
beam width with absence of side lobes and Fresnel zone
which makes the interception of the transmitted signal a
difficult task. Being faster, cheaper and use of unlicensed
1 Introduction spectrum are some of the important features of this type of
communication [8, 9].
The rapid growth of IT and communication based in-
Free space optics (FSO) link being a wireless link faces a
dustries and services in the last few years have given rise to
great deal of problems in communication of information. A
the demand for faster internet speeds and more bandwidth.
narrow beam of light is propagated through free space at-
The number of objects and devices interconnected through
mosphere for transferring the information from transmitting
internet are expected to have a growth from 9.1 billion in
to receiving point which must be in the proper line of sight
the year 2013 to 28.1 billion by the year 2020 [1].This
(LoS) for prevention of any loss of information [10, 11]. The
amounts to be even larger than three times the total ex-
performance of the link is affected by both internal and
pected global population by the year 2020. Wireless
external parameters which can be considered as the main
communication technologies enable user mobility as they
drawback of this communication system. Wavelength, fre-
reach closer to the end user and are thus favored one for
quency, optical power are considered as internal parame-
such systems and applications [2].
ters, on the other hand visibility, scintillation, cloud and fog
Although radio frequency (RF) being a mature tech-
conditions are considered as external parameters [12, 13].
nology has been widely deployed in terrestrial communi-
There may be huge loss in transmission caused by the
cation, the nature of propagation at these frequencies may
variations in weather conditions resulting in reduced vis-
ibility distance. One of the main external parameter that
greatly influences the attenuation characteristics of a FSO
*Corresponding author: Sanmukh Kaur, Amity School of Engineering channel is atmospheric fog conditions. Research on the
and Technology, Amity University, Noida, 201313, Uttar Pradesh,
climate around Graz, Austria suggests that rain may cause
India, E-mail: sanmukhkaur@gmail.com
2 S. Kaur: Performance analysis of FSO link under the effect of fog

a maximum loss of signal up to 25 dB/km while fog may NRZ pulse generator and AM modulator are the compo-
result in the much higher attenuation of up to 100 dB/km in nents of transmitting end whereas PIN photo detector,
the link [14]. Analysis of fog is thus given more priority low pass filter (LPF) and bit error rate (BER) analyzer are
before setting up a FSO link as compared to any other the components at receiving end. The FSO terrestrial
external parameter. model first converts the logical signal to be transmitted
Many researchers have discussed the different im- into electrical signal by using NRZ pulse generator. This
pairments of terrestrial FSO communication along with signal along with CW laser output is applied to the AM
their causes and possible mitigation techniques [7, 15, 16]. modulator which converts the electrical signal into an
M.A. Esmail et al. studied and examined the attenuation intensity modulated optical signal. This resultant signal
modeling of fog attenuation and predicted that FSO will in the form of light is then transmitted by propagating it
have preferred market for 5G network applications with cell into FSO link. The received signal at the other end is in
size of lesser than 1 km diameter [17]. In [18], fog and smoke turn advanced towards the photo detector which converts
attenuation is predicted by proposing a wavelength it into electrical form. BER analyzer at the receiver side
dependent empirical model at visible and near infrared has been used to analyze the quality of the received
(NIR) wavelengths. T. Esmail et al. proposed the compli- signal.
mentary solution of operation of hybrid FSO/mmWave
links as fog has no particular effect on mmWave which are
susceptible to oxygen absorption and heavy rain condi-
tions [14]. In [19] performance of the link has been analyzed
3 Degradation in the performance
using Kim’s model by calculating average attenuation due of the FSO system as a result of
to fog in two winter months.
Delhi is a megacity and second most populated city of
fog induced attenuation
India. The National capital region of Delhi, is a home to
Variation in atmospheric conditions is one of the major
many multinational corporation and giant software devel-
challenges faced by an FSO link. The atmospheric factors
opment companies. The city needs to have better means of
further depend upon time and geographical location of the
communication, which are not affected by weather condi-
particular region. FSO link can experience difference in
tions of the region [20].
attenuation depending upon the location of its setup and
In this work, real time visibility data for three specific
the time at which atmosphere is used as a medium. The
months (January, February and December) of seven
main cause of loss in a FSO link may be atmospheric FOG
consecutive years starting from 2013 to 2019 has been
conditions. The reason why this happens is due to the fact
collected from Delhi, Safdarjung area for calculation of fog
that fog not only absorbs but also scatters light which re-
attenuation. Attenuation data has been obtained from
sults in huge loss of data during transmission in foggy
visibility calculations by applying Beer Lamberts law along
weather. Powerful lasers along with specialized proced-
with Kim and Kruse mathematical models. Received signal
ures have been required to increase the efficiency of
quality has been analyzed as a function of transmitted
transmission.
power, data rate and transmission range considering three
Attenuation is calculated using Beer Lamberts equa-
transmission windows of 850, 1300 and 1550 nm. Further,
tion given as:
the statistical results obtained for two mathematical
−q
models have been analyzed to compare the performance of 3.912 λ
α fog

 (1)
two attenuation models. V(km) 550

Where q represents the scattering size distribution coeffi-

cient, V represents the visibility in kilometer and λ repre-
2 Schematic link configuration sents the wavelength of the transmitted signal. Scattering
size distribution coefficient q may be evaluated by using
Figure 1 gives the schematic representation of terrestrial Kim and Kruse mathematical models. Kruse model may not
FSO link with the addition of bit error rate (BER) analyzer. be able to precisely estimate the specific attenuation if the
The schematic configuration can be divided into trans- visibility lies in the range of <1 km. Kim modified the model
mitting and receiving end, with each end consisting of by applying theoretical assumptions and defined the q
different components. Continuous wave laser or CW laser, values as given in Table 1 [18].
 S. Kaur: Performance analysis of FSO link under the effect of fog 3

Figure 1: Schematic configuration of the FSO

link.

Table : Calculation of q using Kim and Kruse models. communication systems operate in the wavelengths span
of 800–1600 nm [6], three wavelength windows of 850,
Kim model Kruse model 1300 and 1550 nm which are considered nearly transparent
. for V >  km . for V >  km have been chosen for the analysis. Other fixed parameters
. for  km > V >  km . for  km < V <  km have been listed in Table 2.
. + . v for  km > V >  km . V/ for V <  km
We have referred the fog data of Delhi, Safdarjung re-
V–. for . km < V <  km
gion, describing the visibility range in kilometers for each
 for V < . km
day of a year for all regions [21]. From the available data,
average visibility for the three possible fog months
(January, February and December) of seven consecutive
Table : System specifications. years from 2013 to 2019 have been calculated. Tables 3–5
depict the visibility and corresponding attenuation calcu-
Parameter Value lated by using Beer Lamberts law and Kim model for the
Power −– dBm wavelengths of 1550, 1300 and 850 nm respectively.
Wavelength ,  and  nm For the calculated values of attenuations in dB/km as
Bit rate – Gbps depicted in Tables 3–5 at three transmission wavelengths,
Range – km
received signal quality has been analyzed as a function of
Tx aperture diameter . m
Rx aperture diameter . m
transmitted power and propagation range in Figure 2–7.
Modulation scheme NRZ Quality of the received signal has been analyzed at
Beam divergence  mrad 850 nm wavelength in Figures 2 and 3 respectively. As seen
Additional losses  dB in Figure 2, Q-factor at the receiver decreases with increase
Photodetector PIN in propagation distance from 0 to 4 km. Best value of the
Dark current  nA
Q-factor is achieved in the February month in which
transmission upto a range of 3500 m is possible with
acceptable value of BER at the receiver.
4 Results It can be observed form Figure 3 that in the months of
December and January, with corresponding average at-
For improving overall performance and effectiveness of a tenuations of 2.50 and 2.79 dB/km respectively, a power of
FSO link, system parameters are required be optimized for more than 2 dBm should be transmitted as compared to
the location where link efficiency may get reduced as a power in the month of February for the similar quality of
result of fog in the atmosphere. Simulation and analysis of the received signal.
the performance of the system has been accomplished The plots have been obtained as function of trans-
using Optisystem-16 simulation package. For analyzing the mission range and power in Figures 4 and 5 respectively at
performance of the system, some of the internal parameters a wavelength of 1300 nm. It can be observed that the effect
have been set as fixed values while others have been kept of fog induced attenuation loss has been more in the month
as variables. As most of the commercially available FSO of January than the other months. As shown in Figure 4,
4 S. Kaur: Performance analysis of FSO link under the effect of fog

Table : Visibility and attenuation calculation at a wavelength of  nm.

Years January February December

Visibility (km) Attenuation (dB\km) Visibility (km) Attenuation (dB\km) Visibility (km) Attenuation (dB\km)

 . . . . . .

 . . . . . .
 . . . . . .
 . . . . . .
 . . . . . .
 . . . . . .
 . . . .
Average . . .

Table : Visibility and attenuation calculation at a wavelength of  nm.

Years January February December

Visibility (km) Attenuation (dB\km) Visibility (km) Attenuation (dB\km) Visibility (km) Attenuation (dB\km)

 . . . . . .

 . . . . . .
 . . . . . .
 . . . . . .
 . . . . . .
 . . . . . .
 . . . .
Average . . .

Table : Visibility and attenuation calculation at a wavelength of  nm.

Years January February December

Visibility (km) Attenuation (dB\km) Visibility (km) Attenuation (dB\km) Visibility (km) Attenuation (dB\km)

 . . . . . .

 . . . . . .
 . . . . . .
 . . . . . .
 . . . . . .
 . . . . . .
 . . . .
Average . . .

transmission upto a range of 3000 km is possible at this For the calculated values of attenuations in dB/km as
wavelength even for the months of December and January. depicted in Tables 3 at a wavelength of 1550 nm, received
It can be observed form Figure 5 that in the case of signal quality has been analyzed as a function of propa-
operation at 1300 nm transmission wavelength, a power of gation range and transmitted power in Figures 6 and 7
more than 1 dBm is required to be transmitted in the respectively. Similar results have been achieved as in the
December and January months as compared to power used case of operation at 1300 nm wavelength.
in the month of February for the similar quality of the Next we plot the graphs of received signal quality as a
received signal. Results obtained at 1300 nm are better as function of propagation range for different input bit rates
compared to those achieved at the wavelength of 850 nm. ranging from 1 to 10 Gb/s considering three transmission
 S. Kaur: Performance analysis of FSO link under the effect of fog 5

Figure 5: Quality factor versus power at 1300 nm wavelength.

Figure 2: Quality factor versus range for 850 nm wavelength.

Figure 3: Quality factor versus power for 850 nm wavelength. Figure 6: Quality factor versus range at 1550 nm wavelength.

Figure 7: Quality factor versus power at 1550 nm wavelength.

Figure 4: Quality factor versus range at 1300 nm wavelength.

rate and transmission range. The operation at 850 nm

wavelengths. The average attenuation for the three months wavelength does not result in a similar performance and
(January, February and December) of seven consecutive maximum transmission range achieved in the case is also
years from 2013 to 2019 have been computed for obtaining less than 2500 m.
the plots at wavelength of 1550, 1300 and 850 nm shown in Next we perform the descriptive statistical analysis of
Figures 8–10 respectively. refereed visibility data and corresponding optical attenu-
Again we observe that in the case of operation at 1300 ation considering wavelengths of 850, 1300 and 1550 nm
and 1550 nm wavelength (Figures 8 and 9), a similar quality respectively. The attenuation data for both Kim and Kruse
of the received signal is obtained for a given value of bit model has been observed and classified into tables in
6 S. Kaur: Performance analysis of FSO link under the effect of fog

Table 6 lists the statistical analysis parameter values

of the estimated attenuation using Kruse and Kim models
at three different wavelengths for the year 2013. As seen
from the table, there is a slight disagreement in the mean
and median values of the attenuation among the two
models at different wavelengths. Tables 7–12 and Table 13
list the statistical analysis parameter values of the esti-
mated attenuations using two models for the years from
2014 to 2019 and for combined data set respectively. A
similar observation of slight disagreement in the calcu-
lated values of different statistical parameters has been
concluded from the tables. This happens mainly as a
Figure 8: Quality factor versus range at different data rates
considering 1550 nm wavelength.
result of the fact that the observed visibility data has a
small standard deviation which results in the small vari-
ation in the estimated statistical parameters for optical
attenuation using two models.
Further, we compare the statistical results obtained for
two mathematical models at different wavelengths from
the combined data given in Table 13.
Figures 11 and 12 plot the mean, median and range
values of attenuation for three different wavelengths
employing Kruse and Kim model respectively. It can be
observed from the figures that Mean and median values of
attenuation are minimum and maximum at 1550 and
850 nm wavelength respectively. Kim model reports higher
values of these parameters for a given value of wavelength.
Figure 9: Quality factor versus range at different data rates
Both the models result in better performance at 1550 and
considering 1300 nm wavelength. 1300 nm as compared to 850 nm wavelength.

5 Conclusion
In this work, a comprehensive survey of attenuation
caused by different fog conditions of the Delhi, Safdar-
jung region has been reported. From the available fog
data, average visibility and corresponding attenuation for
the three possible fog months (January, February and
December) of seven consecutive years from 2013 to 2019
have been calculated. For the calculated values of atten-
uations at three transmission wavelengths of 850, 1300
and 1550 nm, received signal quality has been analyzed as
Figure 10: Quality factor versus range at different data rates a function of transmitted power, propagation range and
considering 850 nm wavelength. data rate. It has been observed that the effect of fog
induced attenuation is more in the month of January fol-
yearly order for each of the three wavelengths as shown in lowed by December and February months thus requiring
Tables 6–12. an increase in transmitted power in the range of 1–2 dBm
The real time and long term observed visibility data for maintaining the same quality of the signal at the
enables us for estimation of specific value of attenua- receiver.
tion and helps us in determining the availability of the We observe that in the case of operation at 1300 and
link in the region for the complete year as well as whole 1550 nm wavelength a similar quality of the received signal
period. is obtained for a given value of transmission range and
 S. Kaur: Performance analysis of FSO link under the effect of fog 7

Table : Estimated attenuation statistics for the year .

Model Kruse model Kim model

Wavelength  nm  nm  nm  nm  nm  nm

Mean . . . . . .
Median . . . . . .
Range . . . . . .

Table : Estimated attenuation statistics for the year .

Model Kruse model Kim model

Wavelength  nm  nm  nm  nm  nm  nm

Mean . . . . . .
Median . . . . . .
Range . . . . . .

Table : Estimated attenuation statistics for the year .

Model Kruse model Kim model

Wavelength  nm  nm  nm  nm  nm  nm

Mean . . . . . .
Median . . . . . .
Range . . . . . .

Table : Estimated attenuation statistics for the year .

Model Kruse model Kim model

Wavelength  nm  nm  nm  nm  nm  nm

Mean . . . . . .
Median . . . . . .
Range . . . . . .

Table : Estimated attenuation statistics for the year .

Model Kruse model Kim model

Wavelength  nm  nm  nm  nm  nm  nm

Mean . . . . . .
Median . . . . . .
Range . . . . . .
8 S. Kaur: Performance analysis of FSO link under the effect of fog

Table : Estimated attenuation statistics for the year .

Model Kruse model Kim model

Wavelength  nm  nm  nm  nm  nm  nm

Mean . . . . . .
Median . . . . . .
Range . . . . . .

Table : Estimated attenuation statistics for the year .

Model Kruse model Kim model

Wavelength  nm  nm  nm  nm  nm  nm

Mean . . . . . .
Median . . . . . .
Range . . . . . .

Table : Estimated attenuation statistics of combined data of years from  to .

Model Kruse model Kim model

Wavelength  nm  nm  nm  nm  nm  nm

Mean . . . . . .
Median . . . . . .
Range . . . . . .

Figure 11: Mean, median and range statistical parameters at three Figure 12: Mean, median and range statistical parameters at three
different wavelengths using Kruse Model. different wavelengths using Kim Model.

data rate. The operation at 850 nm wavelength does not dictate that Kim model reports higher values mean, median
result in a similar performance and maximum transmission and range statistical parameters for a given wavelength
range achieved in the case is also less than 2500 m. and Kruse model outperforms in terms of quality of the
Results from descriptive statistical analysis of refereed signal received for a given value of range and transmitted
visibility data and corresponding optical attenuation power.
 S. Kaur: Performance analysis of FSO link under the effect of fog 9

Author contribution: All the authors have accepted Processing and Integrated Networks (SPIN). Noida, India; 2019,
responsibility for the entire content of this submitted pp. 552–7. https://doi.org/10.1109/SPIN.2019.8711577.
11. Kaur S, Raza SZA, Khanna J. Analysis of performance of FSO link
manuscript and approved submission.
during the months of monsoon in Delhi, India. In: Jain L,
Research funding: None declared. Tsihrintzis G, Balas V, Sharma D, editors. Data communication
Conflict of interest statement: The authors declare no and networks. Advances in intelligent systems and computing,
conflicts of interest regarding this article. vol 1049. Singapore: Springer; 2020.
12. Rani M, Bhatti H, Singh V. Analysis of atmospheric turbulence on
free space optical system using homotopy perturbation method. J
References Opt Commun 2019;40:441–6.
13. Kaur S. Analysis of inter-satellite free-space optical link
1. International Data Corporation; 2015. [Online]. Available from: performance considering different system parameters. Opto-
https://www.idc.com/. Electron Rev 2019;27:10–13.
2. Hamza AS, Deogun JS, Alexander DR. Classification framework for 14. Ismail T, Leitgeb E, Plank T. Free space optic and mmWave
free space optical communication links and systems. Commun communications: technologies, challenges and applications.
Surv Tut IEEE 2019;21:1346–82. IEICE Trans Commun 2016;E99.B:1243–54.
3. Henniger H, Wilfert O. An introduction to free-space optical 15. Hamza AS, Deogun JS, Alexander DR. Classification framework for
communications. Radio Eng 2010;19:203–12. free space optical communication links and systems. IEEE
4. Hamza AS, Khalifa SS, Hamza HS, Elsayed K. A survey on inter-cell Commun Surv Tut 2019;21:1346–82.
interference coordination techniques in OFDMA-based cellular 16. Mansour A, Mesleh R, Abaza M. New challenges in wireless and
networks. Commun Surv Tut IEEE 2013;15:1642–70. free space optical communications. Optic Laser Eng 2016;89:
5. Kaur N, Goyal R, Rani M. A review on spectral amplitude coding 95–108.
optical code division multiple access. J Opt Commun 2017;38:77–85. 17. Esmail MA, Fathallah H, Alouini MS. Outdoor FSO
6. Bloom S, Korevaar E, Schuster J, Willebrand H. Understanding the communications under fog: attenuation modeling and
performance of free-space optics [Invited]. J Opt Netw 2003;2: performance evaluation. IEEE Photonics J 2016;8:1–22.
178–200. 18. Ijaz M, Ghassemlooy Z, Pesek J, Fiser O, Minh HL, Bentley E.
7. Kaushal H, Kaddoum G. Optical communication in space: Modeling of fog and smoke attenuation in free space optical
challenges and mitigation techniques. Commun Surv Tut IEEE communications link under controlled laboratory conditions. J
2017;19:57–96. Lightwave Technol 2013;31:1720–6.
8. Kaur S, Kakati A. Analysis of free space optics link performance 19. Kesarwani A, Kaur S, Kaur M, Vohra PS. Performance analysis of
considering the effect of different weather conditions and FSO link under diferent conditions of fog in Delhi, India. In: 2018
modulation formats for terrestrial communication. J Opt Commun 2nd IEEE International conference on power electronics,
2018 10 May. https://doi.org/10.1515/joc-2018-0010 [Epub intelligent control and energy systems (ICPEICES). Delhi, India;
ahead of print]. 2018, pp. 958–61.
9. Lath S, Goyal R, Kaler R. A review on free space optics with atmospheric 20. United Nations, Department of Economic and Social Affairs,
and geometrical attenuation. J Opt Commun 2016;37:331–6. Population Division. The World’s Cities in 2016—Data Booklet;
10. Kaur S, Goyal R. Analysis of terrestrial FSO link performance 2016. ST/ESA/SER.A/392.
considering different fog conditions and internal parameters of 21. Support of fog data https://www.timeanddate.com/weather/
the system. In: 2019 6th International Conference on Signal india/new-delhi/historic?month=2&year=2019.