A hybrid algorithm for wave-front corrections applied to satellite-to-ground laser communication

TELKOMNIKA Telecommunication, Computing, Electronics and Control
Vol. 18, No. 3, June 2020, pp. 1259~1267
ISSN: 1693-6930, accredited First Grade by Kemenristekdikti, Decree No: 21/E/KPT/2018
DOI: 10.12928/TELKOMNIKA.v18i3.12960  1259
Journal homepage: http://journal.uad.ac.id/index.php/TELKOMNIKA
A hybrid algorithm for wave-front corrections applied
to satellite-to-ground laser communication
Mohammed Senan Al Gobi1
, Djamel Benatia2
, Mouadh Bali3
1,2
Electronic Department, Laboratoire d'Électronique Avancée (LEA), Faculty of Technology,
University of Batna 2, Algeria
3
Department Computer Science, Faculty of Exact Sciences, Université d’El Oued, Algeria
3
LIMED Laboratory, Faculty of Exact Sciences, Université de Bejaia, Algeria
Article Info ABSTRACT
Article history:
Received Sep 18, 2019
Revised Dec 7, 2019
Accepted Dec 21, 2019
Laser communications hold accurate data rate for ground satellite links.
The laser beam is transmitted through the atmosphere. The clear-air turbulence
induces a number of phase distortions that damage wave-front. Adaptive optics
(AO) treats wave front correction. The nature of AO systems is iterative; it can
be integrated in metaheuristic algorithms such as genetic algorithm (GA).
This paper presents improved version of algorithm for wave-front corrections.
The improved algorithm is based on genetic algorithm (GA) and adaptive
optics approach (OA). It is implemented in a computer simulation model called
object-oriented matlab adaptive optics (OOMAO). The optimisation process
involves best possible GA parameters as a function of population size, iteration
count, and the actuators’ voltage intervals. Results show that the application
of GA improves the performance of AO in wave-front corrections and
the communication between satellite-to-ground laser links as well.
Keywords:
Adaptive optics (AO)
Genetic algorithm
Object-oriented matlab adaptive
optics (OOMAO)
Satellite-to-ground
Wave-front correction This is an open access article under the CC BY-SA license.
Corresponding Author:
Mohammed Senan Al Gobi,
Electronic Department, Laboratoire d'Électronique Avancée (LEA),
Faculty of Technology, University of Batna 2,
Batna, Zip 05000, Algeria.
Email: moh.algobi@gmail.com
1. INTRODUCTION
The communication satellite systems with optical laser links have become priority in communication
fields for number of reasons. Compared to the radio communication which needs more than 1 Gbps [1, 2],
the communication satellite systems with optical laser links data rate is higher (more than 10 Gbps bit rate [3]),
signal intensity (structure of fiber laser [4, 5]) and lower equipment size [5]. It has made a significant
contribution in reducing the effects of atmospheric attenuation that is also determined by the geographic
location especially in the tropical and equatorial regions where the rain effect plays an important role in
the quality of communication [6-8], as it leads to the instability in the intensity and the phase of the received
signals [1, 9]. In laser satellite communications, the elements of the atmosphere (wind, rain, dust…) can affect
the quality of communication [5]. The problem occurs when the optical wave propagates in free space and is
subjected to serious disturbances (wave-front sensor) which renders the system less effective and probably
loses the information.
In adaptive optics (AO), the most important element is the deformable mirror. Which controlled by
the approximation algorithms [9]. There are many algorithms that can correct wave front sensor such as:
the stochastic parallel gradient descent (SPGD), simulated annealing (SA), and genetic algorithm

 ISSN: 1693-6930
TELKOMNIKA Telecommun Comput El Control, Vol. 18, No. 3, June 2020: 1259 - 1267
1260
(GA) [10, 11]. This contribution provides a hybrid solution to correct wave front sensor. The provided solution
consists of the combination of GA with AO solution. The hybrid solution gives positive results in correcting
the wave front aberration in satellite laser communication.
The present paper is divided into five sections. A brief overview of the effects of atmosphere
attenuation on the refractive index's behaviour is presented in section two. In section 3 provides the design and
implementation of the genetic algorithm (GA) and the hybrid (GA and AO). Section 4 is dedicated to
the discussion of the obtained results. The last section is devoted to the overall results of our study and
future works.
2. BACKGROUND OF THE STUDY
When the optical signal (laser here) passes through the atmosphere, it gets exposed to atmospheric
attenuations i.e. clouds, rains, winds, dust. As a result, the satellite-to-ground optical links will be degraded.
Figure 1 illustrates the overall frequency behavior of the real part of the index of refraction. Atmospheric
attenuations is the aberration in wave-front deformation. It reduces the intensity and the phase of the received
signals. As result, it decreases the image's resolution. To correct phase aberration, wave-front correction is
implemented using adaptive optics (AO) technology [5, 9].
Figure 1. Behavior of the index of refraction as a function of linear frequency [12]
2.1. The theory of atmospheric fluctuations
Atmospheric fluctuation at Ka-band frequencies is commonly known as atmospheric turbulence. Each
point in the telescope aperture receiving is considered as the sum of many components distributed by
the turbulent eddies. Constructive interference of laser waves appear as spots of light. These spots represent
the waves interfered constructively. However, destructive interference appears as dark area. The effects of
atmospheric turbulence are refined by changing the telescope aperture and by reducing the amplitude
scintillations observed at the receivers [13].
2.1.1. Classical atmospheric turbulence
Two extreme turbulent scales (the internal l0, and the external scale L0) have been identified to model
the atmospheric turbulence [5, 8]. The internal scale l0 (r, t) corresponds to the spatial scale from which
the kinetic energy is dissipated in heat by viscous friction. Therefore, it depends on the density of
the atmosphere. l0 (r, t) can vary from a few millimetres near the ground to a few centimeters in
the tropopause (interface between the troposphere and the stratosphere) [9]. The external scale L0 (r, t) (ranging
from 10-100 meters) is determined by the size of the of air masses. It corresponds to the largest macroscopic
phenomena (air layers, winds, weather disturbances). The inertial domain defines the scales for which
turbulence is fully developed. It determines limiting value of the external scale of turbulence L0 and
the internal scale of turbulence l0 [9]. In a turbulent air, the random variations of refractive index, n, are
described by the structure function, Dn [13].
𝐷 𝑛(𝜌) = ⟨|𝑛(𝑟) − 𝑛(𝑟 + 𝜌)|2⟩ (1)
The difference between the value n (r) at a point r and the value n (r +ρ) at a point (r +ρ) is a point
distant from the vector r in a distance ρ. It represents a position in a three-dimensional space [14]. For
an established turbulent regime in the inertial domain, the variance of the difference at two points of space, or
structure function, is given by:
𝐷 𝑛 = {
𝐶 𝑛
2
𝜌
2
3⁄
, 1
𝑙0
⁄ ≤ 𝜌 ≤ 1
𝐿0
⁄
𝐶 𝑛
2
𝑙0
2
3⁄
(
𝜌
𝑙0
)
2
, 𝜌 < 1
𝑙0
⁄
(2)

TELKOMNIKA Telecommun Comput El Control 
A hybrid algorithm for wave-front corrections applied to… (Mohammed Senan Al Gobi)
1261
where 𝐶 𝑛
2
is the structure function constant expressed in m-2/3
and displacement (ρ = | ρ |) is a scalar measured
by m. 𝐶 𝑛
2
is typically vary between 10-13
m-2/3
and 10-15
m-2/3
for strong turbulence and weak turbulence
respectively [5, 15].
2.1.2. Spectral density and variance of refractive index
According to the Wiener-Khinchine theorem, the power spectral of the spatial fluctuations is
calculated using simple Fourier transform [14]:
∅ 𝑛(𝑘) = 0.033𝐶 𝑛
2
𝑘
−11
3⁄
(3)
𝑘: is the modulus of the spatial frequency (Kolmogorov spectrum) that is expressed in m-1
.
The Kolmogorov spectrum (3) is valid only in the inertial domain: 1/L0 < k < l/10. It represents an external
scale and an internal scale of turbulence respectively [14].
The variance of intensity fluctuations 𝜎 𝜒
2(𝑑𝐵2) at the terminal point of the receiver after an (initially)
plane wave with wave-number k=2𝜋/𝜆 has propagated through a depth L of homogeneous turbulence
is [13, 15]:
𝜎𝜒
2
= {
186 𝑙0
−7
3⁄
𝐶 𝑛
2
𝐿3
, 𝑙0 > √𝜆𝐿
23.2 𝐶 𝑛
2
𝑘
7
6⁄
𝐿
11
6⁄
, 𝐿0 ≪ √𝜆𝐿 < 𝑙0
75.4 〈(∆𝑛)2
〉𝐿𝐿 𝑛 𝑘2
, 𝐿0 ≪ √𝜆𝐿
(4)
where 〈(∆𝑛)2〉: is square refractive index fluctuation and Ln is the integral scale of the turbulence of
the same order as L0. The first expression on the right side of (4) refers to the optical regime whereas
the second expression is the diffraction regime. The latter plays an important role in microwave scintillation
that occurs on Earth–space paths. At the terminal point of the receiver, the expected spectral density, 𝑊𝜒( 𝜔)
of the signal is flat at a low frequency and rolls off at high frequencies. The asymptotic behaviour is given
by [15, 16].
𝑊𝜒(𝜔) → {
2.765
𝜎 𝜒
2
𝜔 𝑡
, 𝜔 → 0
7.13
𝜎 𝜒
2
𝜔 𝑡
(
𝜔
𝜔 𝑡
)
−8
3⁄
, 𝜔 → ∞
(5)
where ωt = νt (k/L)1/2
is the Fresnel frequency and νt is the component of the wind velocity transverse to
the propagation path. The two asymptotes intersect at a corner frequency, ωc, which depends on νt. For a thick
layer of uniform (homogeneous) turbulence ωc =1.43 ωt [13].
2.2. Principal of adaptive optics
Adaptive optics (AO) technology helps correcting the phase aberration in the laser wave-front caused
by atmosphere turbulence in real time. AO is used in space field and in a number of fields such
as: ophthalmology [17]. Recently the adaptive optics is also used in the two-photon excitation microscopy [17].
To correct the wave-front in the optical systems, AO systems are used with a deformable mirror
(DM) [10, 17]. In our study, optimization algorithms individually controls The 92 actuators' deformable mirror.
They put each actuator in the right position by changing the mirror surface shape to produces a suitable wave
front for the system [11].
2.2.1. Wave-front analysis
Unlike in the radio frequency field, it is not yet possible to measure the phase of the wave-front at
optical wavelengths directly. Optical detector is incapable to respond to temporal frequencies. This problem is
overcome by performing indirect measurements i.e. by analyzing the impact of phase disturbances on
the intensity distribution [9]. Rousset carries out a description to analyzers for adaptive optics [18].
The Shack-Harmann analyser is commonly used in adaptive optics because its limitations are
the representatives of those of most planar pupil analyzers [14].
2.2.2. Principal of the Shack-Hartmann analyzer
The Shack-Hartmann wave surface analyzer (SH) is a pupil plane analyzer as shown in Figure 2.
It is based on the geometrical optics formalism [19]. A micro-lens array samples the incident wave-front in
the pupil plane. A measurement of the position of the image spot formed at the focus of each of
the micro-lenses gives access to the local slope of the wave-front in the pupil plane of each micro-lens [14].
The measurement of the position is often carried out by the center of gravity, but other positional estimators

 ISSN: 1693-6930
1262
can also be used, such as correlation [19-21]. The slope measured by the center of gravity in each sub-pupil k
is respectively for the direction in x and y:
𝑃𝑥
𝑘
=
∬ 𝑑𝑟
𝛿𝜑 𝑘(𝑟)
𝛿𝑥
|𝜓 𝑘(𝑟)|2
∬ 𝑑𝑟|𝜓 𝑘(𝑟)|2 (6)
𝑃𝑦
𝑘
=
∬ 𝑑𝑟
𝛿𝜑 𝑘(𝑟)
𝛿𝑦
|𝜓 𝑘(𝑟)|2
∬ 𝑑𝑟|𝜓 𝑘(𝑟)|2 (7)
where the double integration is carried out on the surface of the sub-pupil k considered, φk the phase and | ψk |
the amplitude of the complex field. When the intensity is constant in each sub-pupil, the slope measurement is
then an average on the surface of the sub-pupil.
Figure 2. Schema of an adaptive optics system and the Shack-Hartmann wave-front analyser
2.3. Genetic algorithm
Genetic algorithm (GA) is a stochastic parallel algorithm that uses Darwin's theory of species
evolution. It is based on three principles: variation, adaptation and inheritance [22, 23]. Variation: every
individual within a population is different. The differences, more or less important, will be decisive in
the process' selection. Adaptation: The most adapted individuals to their environment reach adulthood more
comfortably. Individuals with better survivability will therefore be able to reproduce more. Heredity:
The characteristics of individuals must be hereditary in order to be transmitted to their descendants.
This mechanism will make it possible to evolve the species to share the advantageous characteristics to its
survival [23-25]. In the present work, population refers to the possible collection of solutions. However,
the individual represents a solution. Chromosome, in the other hand, is a component of the solution and gene
represents a characteristic (or a peculiarity).
2.4. Object-oriented matlab adaptive optics (OOMAO)
Object-Oriented Matlab Adaptive Optics (OOMAO) is a library of Matlab classes, dedicated to
adaptive optics (AO) systems. The main classes used in this toolbox are: source, atmosphere, telescope,
Shack-Hartmann, deformable mirror [26, 27]. The source class is the link between other classes.
The atmosphere class contains all the parameters defining the atmosphere. A multilayer atmosphere is created
by setting the appropriate vectors of altitudes, wind speeds and directions and turbulence strengths.
An atmosphere object needs to be coupled with a telescope object to create a 3-D volume of turbulence phase
screens. The telescope class contain the telescope parameters and the phase screens in the turbulent layers set
by an atmosphere object. In a closed-loop adaptive optics system, the deformable mirror is the first active
component that encounters the wave front. To complete an Adaptive Optics System, there must be a wave front
sensor. The OOMAO implements the wave front sensor [26, 28].

1263
3. OUR PROPOSED SLOUTION
In this section, we discuss the design of a hybrid solution based on OOMAO closed-loop with
the assistance of the genetic algorithm.
3.1. Genetic algorithm notation
− Population: matrix Pop(NxM) of real numbers contains the set of solutions generated by one iteration as
shown in Table 1: where N is the number of solutions (or chromosomes) and M is the number of actuators
in DM (or genes).
− Chromi: vector of real numbers contains the solution (i) in the population.
− Genej: real number, represents the value of the assigned voltage to the actuator j in the DM
Table 1. Population coding
I x J Gene1 Gene2 …. Genej
Chrom1 0.08 0.0025 … -0.036
Chrom2 …
... …
… …
Chromi …
3.2. Hybrid algorithm flow-chart
The main steps of the processes in our Algorithm are:
− Initialization : values are assigned to GA parameters (population size, iterations count…), and generate
the chromosomes by assigning random values to their genes.
− Evaluation : fitness function is defined to determine the adaptation score of chromosomes during
the selection process. We define the fitness (i) for chromi.
𝑭𝒊𝒕𝒏𝒆𝒔𝒔(𝒊) =
∑ 𝝈𝒊𝒋
𝑴
𝒋=𝟏
𝑴
(8)
where 𝛔ij is the standard deviation for Gaussian mutations of the sensor j, and M is the number of the actuators.
− Selection: according to the roulette method of selection, we select a set of best chromosomes, will be
referred to as parents, in order to reproduce the next generation. The Figure 3 illustrates the main steps of
the processes in hybrid algorithm solution.
Figure 3. Flow-chart of the hybrid solution

 ISSN: 1693-6930
1264
4. RESULTS AND DISCUSSION
Our experimental work has passed into two steps. Firstly, we looked for the best configuration for
genetic algorithm parameters (population size-iteration count-actuator current interval), in order to get the best
performance of the genetic algorithm. The second step is the execution of the OOMAO closed loop [28] and
let the genetic algorithm use the obtained results in its initial population (as 50% of individuals). Also, it uses
the best parameters from the first step. Then, we made a comparison between OOMAO closed loop and our
hybrid algorithm.
4.1. AO parameters
For AO parameters, data are amassed by using OOMAO founders in [26, 28]. We present them
in Table 2.
Table 2. AO parameters as used in this analyse
Parameter Value
Altitude [0, 4000, 10000]
Atmosphere fractionnalR0 [0.7, 0.25, 0.05]
Wind speed [5, 10, 20]
Wind Direction [0, pi/4, pi]
Source Wavelength 60
Shack Hartmann (wave sensor) Wave front sensing 700nm
lens let array 92
Camera 542 pixels
Diameter 8m
Telescope Resolution 54
Field of View 2.5 arc minute
Sampling Time 500Hz
4.2. GA parameter
Using the data shown in Table 2, we made more than 60 experiments in differents situations
using a number of GA paramaters such as: population size, Iterations number and Min (and Max) value of
voltage. To obtain the best GA parameters, we plot the root mean square values in µm of the wave-front
corrections as a function of population size, iteration accounts, and voltage intervals. Results are shown in
Figures 4, 5, and 6 respectively. Table 3 summarizes the GA parameters as obtained from this study.
Figure 4 represents the Fitness according to the size of the population. It is noted that the Fitness varies
inversely as a function of the size of the population, as the figure shows; in this case we can see Fitness
fluctuations for small number of population size, then the reduction of Fitness is obvious as the size of
the population increases until a certain value of the population size where the Fitness remains steady when
the population size increases.
Figure 5 shows the variations of the fitness according to the number of iterations. From this curve, we
can see that fitness varies enormously with number of iterations. For variable iteration count values from 30 to
4000 iterations, Fitness overall trend is downward, fallen to about 0.4439 μm. We notice in this case how
the number of iteration is significantly effects fitness more than the size of population does. Figure 6 shows
the variations of the fitness according to the voltage intervals of the actuators. This histogram indicates that
the best range (interval) of the current is variable from [-1e-7
to + 1e-7
] to obtain the best Fitness varies less than
0.61 μm. To conclude, the best configurations for the GA parameters that lead to the best wave-front corrections
are those who correspond to population size ≥ 2000, iteration count ≥ 2500, and a voltage interval (-1e-7
…
1e-7
).
Table 3. Gentic algorithm parameters
Value Description
Min-coef -1e-7
Minimum value of voltage to be applied on DM actuators
Max-coef +1e-7
Maximum value of voltage to be applied on DM actuators
popSize 2500 Population size (Chromosomes number)
Chrom Size 80 Chromosome size (number of genes/actuators)
Iterations Count >3000 Genetic loops count
Cross Rate 60% Crossover rate
Cross Points Count 40 Crossover points count
mutRate 70% Mutation rate

1265
Figure 4. Wave-front correction as function of popsize
Figure 5. Wave-front correction as a function of iteration counts
Figure 6. Histograms of wave-front corrections as a function of the actuator voltage intervals

 ISSN: 1693-6930
1266
4.3. Comparison between OOMAO and hybrid algorithm
We propose a hybrid solution (AO and GA). When the initial population of our GA is not totally
random, we took some real solutions from the default OOMAO closed loop (50% and the rest of individuals
are random), then we compare the results. Figure 7 represents results obtained from the comparison between
the OOMAO which proposed in [26, 28] and our hybrid solution. It can clearly be seen that there has been
a sharp decrease in both graphs between 0 and 200 iteration because 50% of suggested solutions are random.
After 300, hybrid solution has only shown a slight growth. This is because algorithm excludes the wrong
solutions and start keeping only the correct ones. After 200, the graph has a horizontal pitch. Between
(2500-3000), there is a stability in hybrid solution graph which continues to drop making a better performance
after 3000 iteration.
Overall, the graph illustrates the comparison between two solutions. Firstly, we notice that
the values of hybrid solution is bigger than OA solution. That’s because the number of iterations is small (less
than 200) which leads to very weak result. Then, the number of iterations grows up making a better results
because our GA is getting optimized by the time.
Figure 7. Comparison between OOMAO solution and the proposed hybrid solution
5. CONCLUSION
The genetic algorithm was applied in adaptive optical system to correct wave-front sensor in
a disturbed atmosphere. Based on the OOMAO, we simulate the performance of an adaptive optical system
with telescope (resolution 54, diameter 8m) and deformable mirror containing 92 actuators. We used the best
parameters of the previous simulation (popSize 2500, Iterations count > 3000, min-volt -1e-7
, max-volt +1e-7
)
to create hybrid solution (AO and GA). By comparing the obtained results, it was found that using this hybrid
solution leads to better enhancement of AO performance than those produced by applying the OOMAO with
the difference of 0.0279µm. We intend in future works to reduce the response time of the GA and develop
more optimization algorithms particularly those related to the improvement of the wave-front sensor like:
simulated annealing (SA), algorithm of pattern extraction (Alopex) and stochastic parallel gradient
descent (SPGD).
REFERENCES
[1] Zhu X., Kahn J. M., “Free-space optical communication through atmospheric turbulence channels,” IEEE
Transactions on Communications, vol. 50, no. 8, pp. 1293-1300, August 2002.
[2] Ippolito L. J., “Radiowave propagation in satellite communications,” Springer Science & Business Media, 2012.
[3] Arimoto Y., Hayano Y., Klaus W., “High-speed optical feeder-link system using adaptive optics,” Proceedings of
SPIE - The International Society for Optical Engineering, January 1997.
[4] Siles G. A., Riera J. M., Garcia-del-Pino P., “Atmospheric attenuation in wireless communication systems at
millimeter and THz frequencies [Wireless Corner],” IEEE Antennas and Propagation Magazine, vol. 57, no. 1,
pp. 48-61, February 2015.
[5] Barchers J. D., Fried D. L., “Optimal control of laser beams for propagation through a turbulent medium,” Journal
of the Optical Society of America A (OSA), vol. 19, no. 9, pp. 1779-1793, 2002.

1267
[6] Abozeed M. I., Alhilali M., Yin L. H., Din J., “Rain attenuation statistics for mobile satellite communications
estimated from radar measurements in Malaysia,” TELKOMNIKA Telecommunication Computing Electronics and
Control, vol. 17, no. 3, pp. 1110-1117, June 2019.
[7] Alhilali M., Lam H., Din J., “Comparison of Raindrop Size Distribution Characteristics across the Southeast Asia
Region,” TELKOMNIKA Telecommunication Computing Electronics and Control, vol. 16, no. 6, pp. 2522-2527,
December 2018.
[8] Badron K., Ismail A. F., Din J., Tharek A. R., “Rain induced attenuation studies for V-band satellite communication
in tropical region,” Journal of Atmospheric and Solar-Terrestrial Physics, vol. 73, no. 5-6, pp. 601-610, April 2011.
[9] Yang H., Li X., “Comparison of several stochastic parallel optimization algorithms for adaptive optics system without
a wavefront sensor,” Optics & Laser Technology, vol. 43, no. 3, pp. 630-635, April 2011.
[10] Chen E., Cheng H., An Y., Li X., “The Improvement of SPGD Algorithm Convergence in Satellite-to-Ground Laser
Communication Links,” Procedia Engineering, vol. 29, pp. 409-414, 2012.
[11] Avanaki M. R., Hojjatoleslami S., Sarmadi H., Ebrahimpour R., Podoleanu A. G., eds., “Genetic algorithm for
optimization of optical systems,” 2010 18th
Iranian Conference on Electrical Engineering, pp. 172-176, 2010.
[12] Jackson J. D., “Classical electrodynamics john wiley & sons,” Inc, New York, pp. 832, 1999.
[13] Yu P., Glover I. A., Watson P., Davies O., Ventouras S., Wrench C., “Review and comparison of tropospheric
scintillation prediction models for satellite communications,” International Journal of Satellite Communications And
Networking, vol. 24, no. 4, May 2006.
[14] Voyez J., “Mesures optiques de profils de turbulence atmosphérique pour les futurs systèmes d'optique adaptative,”
PhD Thesis, Université Nice Sophia Antipolis, 2013.
[15] Kolmogorov A. N., “The local structure of turbulence in incompressible viscous fluid for very large Reynolds
numbers,” Proceedings of the Royal Society of London Series A: Mathematical and Physical Sciences, vol. 434,
no. 1890, pp. 9-13, July 1991.
[16] Brussard G. and Watson P. A., “Atmospheric modelling and millimetre wave propagation,” Springer
Netherlands, 1994.
[17] Cha J-W., Ballesta J., So P. T., “Shack-Hartmann wavefront-sensor-based adaptive optics system for multiphoton
microscopy,” Journal of biomedical optics, vol. 15, no. 4, July 2010.
[18] Rousset G., “Wave-front sensors,” Adaptive optics in astronomy, 1999.
[19] Poyneer L. A., “Scene-based Shack-Hartmann wave-front sensing: analysis and simulation,” Applied Optics, vol. 42,
no. 29, pp. 5807-5815, 2003.
[20] Rais M., Morel J-M., Thiebaut C., Delvit J-M., Facciolo G., “Improving wavefront sensing with a Shack-Hartmann
device,” Applied Optics, vol. 55, no. 28, pp. 7836-7846, 2016.
[21] Weyrauch T., Vorontsov M. A., Bifano T. G., Hammer J. A., Cohen M., Cauwenberghs G., “Microscale adaptive
optics: wave-front control with a µ-mirror array and a VLSI stochastic gradient descent controller,” Applied Optics,
vol. 40, no. 24, pp. 4243-4253, 2001.
[22] Grefenstette J. J., “Genetic algorithms and their applications,” proceedings of the second international conference on
genetic algorithms: Psychology Press, July 1987.
[23] Chambers L. D., “The practical handbook of genetic algorithms,” applications: Chapman and Hall/CRC, pp. 544,
December 2000.
[24] Nosato H., Itatani T., Murakawa M., Higuchi T., Noguchi H., “Automatic wave-front correction of a femtosecond
laser using genetic algorithm,” 2004 IEEE International Conference on Systems, Man and Cybernetics (IEEE Cat.
No.04CH37583), The Hague, vol. 4, pp. 3675-3679, 2004.
[25] Yang P., Hu S., Chen S., Yang W., Xu B., Jiang W., “Research on the phase aberration correction with a deformable
mirror controlled by a genetic algorithm,” Journal of Physics: Conference Series, vol. 48, no. 1, pp. 1017-1024,
October 2006.
[26] Conan R., Correia C., “Object-oriented Matlab adaptive optics toolbox,” in Proceedings of SPIE-The International
Society for Optical Engineering, pp. 9148-91486C, August 2014.
[27] Chulani H. M., Rodríguez-Ramos J. M., “Preliminary performance results of the weighted Fourier phase slope
centroiding method for Shack–Hartmann wavefront sensors obtained with the OOMAO simulator,” Conference:
Adaptive Optics for Extremely Large Telescopes V (AO4ELT5), June 2017.
[28] Github E., “Ethereum rconan/OOMAO. Object-Oriented, Matlab & Adaptive Optics,” 2018. [Online]. Available:
https://github.com/rconan/OOMAO. 2018

A hybrid algorithm for wave-front corrections applied to satellite-to-ground laser communication

More Related Content

What's hot (18)

Similar to A hybrid algorithm for wave-front corrections applied to satellite-to-ground laser communication (20)

More from TELKOMNIKA JOURNAL (20)

Recently uploaded (20)

A hybrid algorithm for wave-front corrections applied to satellite-to-ground laser communication