rus
Issue #5/2020
M. Yu. Kernosov, S. N. Kuznetsov, B. I. Ognev, A. A. Parshin
Reduction of the Level of Errors in the Transmission of High-Frequency Optical Signals in a Turbulent Atmosphere Due to the Use of Statistics of the Received Signal Level
Reduction of the Level of Errors in the Transmission of High-Frequency Optical Signals in a Turbulent Atmosphere Due to the Use of Statistics of the Received Signal Level
DOI: 10.22184/1993-7296.FRos.2020.14.5.424.436
Analysis of the distribution probability of the received signal in a wireless optical communication line under various conditions shows a significant dependence of the error level in the communication channel on the level of power stabilization at the receiver. The significant difference between the statistics of the received signal level and the known distributions does not allow using them to approximate the probabilities of large deviations of the signal level from the mean value. An algorithm for evaluating the optimality and adjusting the level of stabilization of the receiving power based on the distribution statistics is proposed. According to experimental data, the proposed algorithm significantly reduces the level of errors in the atmospheric communication channel.
Analysis of the distribution probability of the received signal in a wireless optical communication line under various conditions shows a significant dependence of the error level in the communication channel on the level of power stabilization at the receiver. The significant difference between the statistics of the received signal level and the known distributions does not allow using them to approximate the probabilities of large deviations of the signal level from the mean value. An algorithm for evaluating the optimality and adjusting the level of stabilization of the receiving power based on the distribution statistics is proposed. According to experimental data, the proposed algorithm significantly reduces the level of errors in the atmospheric communication channel.
Теги: atmospheric optical communication lines errors in a wireless communication channel free-space optics laser communication terminal power stabilization at an optical receiver wireless optical communications атмосферные оптические линии связи беспроводные оптические коммуникации ошибки в беспроводном канале связи стабилизация мощности на оптическом приемнике
Reduction of the Level of Errors in the Transmission of High-Frequency Optical Signals in a Turbulent Atmosphere Due to the Use of Statistics of the Received Signal Level
M. Yu. Kernosov, S. N. Kuznetsov, B. I. Ognev, A. A. Parshin
Mostcom JSC, Ryazan, Russia
Analysis of the distribution probability of the received signal in a wireless optical communication line under various conditions shows a significant dependence of the error level in the communication channel on the level of power stabilization at the receiver. The significant difference between the statistics of the received signal level and the known distributions does not allow using them to approximate the probabilities of large deviations of the signal level from the mean value. An algorithm for evaluating the optimality and adjusting the level of stabilization of the receiving power based on the distribution statistics is proposed. According to experimental data, the proposed algorithm significantly reduces the level of errors in the atmospheric communication channel.
Keywords: laser communication terminal, free-space optics, atmospheric optical communication lines, wireless optical communications, power stabilization at an optical receiver, errors in a wireless communication channel
Received on: 10.07.2020
Accepted on: 24.08.2020
INTRODUCTION
The active development of wireless optical communications, which have certain advantages over other wireless solutions, in particular, is towards an increase in the data transmission rate [1, 2]. Optical wireless solutions have recently become especially relevant in high-speed communications of low-flying (LEO) spacecraft with the Earth [3–5]. The increase in the data transfer rate is unambiguously associated with a decrease in the dynamic range of the receiving devices. On the other hand, the dynamic range of the receiver directly determines the level of errors caused by atmospheric turbulence [6, 7]. In this regard, the optimization of the parameters of the transmit / receive channel becomes especially important.
In this paper, the study of optimization possibilities was carried out using Artolink wireless optical communication equipment designed to operate at speeds of 10 Gbps [8].
PROBLEM STATEMENT
One of the promising directions in the development of wireless optical communication systems is the creation of a wireless channel for transparent connection of high-speed interfaces of user equipment. This makes it easy to scale the wireless optical solution in the direction of increasing the transmission speed [9, 10]. On the other hand, a decrease in the dynamic range of the receiver, accompanying an increase in the transmission rate (mainly from the side of the lower boundary of the received powers – sensitivity), taking into account the influence of the atmosphere, requires the selection of the optimal value of the average signal level at the reception from the atmosphere, the stabilization point (SP). Typically, at 10 Gbps, the receiver (based on the PIN diode) has a dynamic range of about 21 dB (at a sensitivity of –18 dBm and a saturation power of 3 dBm). Taking into account the possible level of oscillations of the receiving signal in the atmosphere in a range comparable to the dynamic range of the receiver [11], the choice of the SP determines the level of digital errors in the channel. For a reasonable choice of the SP, it is necessary to know the statistics of the values of the received power (Iin) in order to minimize the number of events when Iin goes beyond the dynamic range of the receiver.
MODEL ASSESSMENT
At the first stage, a model assessment of the level of possible errors was carried out for various levels of signal instability and SP values. According to [12], the probability function of the receiving signal level when the laser radiation passes through the atmosphere is close to the lognormal law. Therefore, for a model assessment, the level of possible errors was calculated for a lognormal distribution.
The probability density of the lognormal distribution of the received power level relative to the mean value is described by the expression:
, (1)
where In is the received radiation power, normalized to the average value; σn is the root-mean-square deviation of the received radiation power, referred to its average value.
The value σn serving as a measure of the instability of the receiving signal is determined by the formula:
, (2)
where i is the ordinal number of the measurement in the sample; Iini is the value of the received power in the i-th dimension; Iavg the average value of the received power; N is the number of values in the sample.
The error level Perr was determined as the sum of the error probabilities due to a decrease in the signal level below the sensitivity level and an excess of the receiver saturation level according to the formula:
, (3)
where PSens is the probability of events with Irec below the receiver sensitivity level; PSat the probability of events with Irec above the saturation level of the receiver.
PSens and PSat were calculated using the following formulas:
(4)
(5)
where Imin is the receiver sensitivity level; Imax is the receiver saturation power level.
The calculations were performed for the σn levels at which Perr takes values from 10–16 to 10–2. The target range of Perr was reached at σn values from 0.1 to 0.5.
Taking into account that the middle of the receiver’s dynamic range corresponds to –6 dBm, the SP was set in the range from –10 to –2 dBm, corresponding to input signal levels from 100 to 600 μW. The dependences of Perr on SP for various values are shown in Fig. 1.
As seen from Fig. 1, the model dependence of the SP has an optimal value from the point of view of minimizing losses. The optimal SP value is in the region of –5 dBm, which differs from the midpoint of the receiver dynamic range of –6 dBm. The Perr level for optimal SP values is highly dependent on the level of signal instability, which is mainly determined by the distance and weather conditions on the track.
METHOD AND CONDITIONS OF MEASUREMENT
For measurements, a set of Artolink M1–10GE equipment was used, consisting of two terminals, each of which includes an optical unit (OU) – a transceiver, an interface unit (IU) and a connecting cable.
The software of the equipment allows to change the SP value, which is automatically maintained at the receiver of each terminal by adjusting the output power of the optical amplifier (OA) of the opposite transceiver using the service channel between the terminals. The Iin values with a frequency of 100 Hz were recorded on a computer connected to the interface for monitoring the state of the equipment on the IU.
The optoelectronic diagram of the transmitting and receiving path is shown in Fig. 2.
According to [13], the main factors causing fluctuations in the signal level at the reception are atmospheric turbulence and aperture limitation of the beam in the receiving plane. To analyze the statistics of Irec, the following series of measurements were carried out:
at a distance of 620 m at night and in the daytime;
at a distance of 620 m at night and in the daytime with aperture limitation (20 mm diaphragms – due to radiation divergence, this corresponds to fully opened apertures at a distance of 3,000 m) at one terminal corresponding to a distance of about 3,000 m;
at a distance of 1,600 m in the daytime;
at a distance of 2,800 m in the daytime.
In order to avoid the influence of changes in weather conditions during the measurements, each session was recorded for 10 minutes, which made it possible to record about 130,000 Iin values.
It should be noted that during all measurement sessions the following subsystems of the equipment worked: autotracking; adjustment of the direction of communication to the maximum of the received signal on each transceiver; stabilization of the received signal level by adjusting the power of the optical amplifier. Thus, the measurements were carried out under the conditions of fully functional operation of the equipment, which, in fact, was necessary, since the purpose of the work was to analyze and optimize the working version of the equipment.
MEASUREMENT RESULTS
“Fig. 3 shows the Iin values dynamics in μW for a SP of 350 μW at distances of 620, 1,600 and 2,800 m. As it can be seen, the depth of the falls of the values of Iin relative to the SP grows with an increase in the distance and reaches 10 dB at 2,800 m towards increasing its values.
To determine the actual dependence of the error level in the transmission channel on the SP value at a distance of 2,800 m, measurements were made of the dependence of the bit error level (BER) in the communication channel on the SP value. The measurement results are shown in Fig. 4.
It is clear that the SP clearly has an optimal value in terms of minimizing losses. In addition, a change in SP by only 50 μW leads to a change in BER level by an order of magnitude, which indicates the relevance of the SP optimization problem.
The choice of the optimal SP value depends on the nature of the distribution statistics of Iin values, the form of which, according to [14], significantly affects the BER level.
To quantify the nature of the statistics of the distribution of In values for all samples of In values obtained in the series of measurements, the values of the normalized root-mean square deviation σn and the asymmetry coefficient (Ca) of the distribution of the sample were calculated. The calculation of the σn value was carried out according to the formula (2). The following formula was used to calculate Ca:
. (6)
The results of calculating the parameters of statistics for various conditions are shown in the table. Their results show the following:
the instability of the receiving signal significantly depends on the turbulence of the atmosphere, increasing in the morning and decreasing at night;
σn increases with increasing distance, which, as shown in [9], is mainly determined by an increase in the aperture limitation of the receiving beam;
Ca in all series of measurements is nonzero and sometimes changes sign when going from a linear scale Iin to a logarithmic one.
Nonzero values of Ca and a change in sign indicate an intermediate character of the distribution density function In. Thus, with a small effect of atmospheric turbulence and aperture limitation, the distribution is closer to normal, and to lognormal with an increase in these effects.
Thus, the measurement results showed that a qualitative approximation of the real probability distribution by known types of distributions is not possible. In addition, one should take into account the fact that the integrals of the peripheral parts of the probability function play the main role in estimating the errors of the communication channel. In this regard, to solve the problem of optimization of the SP, a methods for assessing the peripheral parts of the distribution was required, which in the equipment should work in real time.
A solution was proposed that makes it possible to evaluate the peripheral parts of the specified function, since the error level is completely determined by the fraction of events of Irec going beyond the receiver’s dynamic range. Thus, SP tuning should minimize the number of these cases. Due to the fact that the measurement frequency of Iin in the equipment is 100 Hz, at an error level of 10–9, it will take 107 s to ensure that the dynamic range is exceeded, which corresponds to a period of about 2,800 hours.
To obtain a reasonable tuning period for the SP, it is sufficient to determine the statistics of achieving Iin at a certain distance from the boundaries of the dynamic range. However, due to their remoteness from the boundaries of the dynamic range, to take into account the shape of the “tails” of the probability function (PF), it is necessary to take into account their shape. For this, a method was proposed for assessing the peripheral parts of the PF based on the “statistical pattern” of the channel – an experimentally determined statistical image of the edges of the probability distribution Iin. To calculate the pattern, 4 parameters were used. They represent fractions of measurements with Iin values lying below and above certain boundaries, respectively (located at a certain distance from the boundaries of the receiver’s dynamic range). Fig. 5 shows an example of the experimental distribution of the number of measurements N depending on Iin, as well as the boundaries for calculating the Iin values to determine the statistical pattern of the channel.
As the lower boundaries of LowOUT and LowIN, the Pin values were chosen, which exceeded the receiver sensitivity by 3 and 6 dB (2 and 4 times), respectively. As the upper boundaries of HighOUT and HighIN – Iin, lower saturation powers of the receiver by 3 and 6 dB, respectively.
To test the proposed method for adjusting the SP, an algorithm for changing the SP was implemented depending on the result of the analysis of the statistical pattern obtained in 1 minute of equipment operation.
Statistics are collected for N = 30,000 receive power measurements using the following counters:
LowIN – power meter is less than the inner minimum limit;
LowOUT – power meter is less than the outer minimum limit;
HighIN – power measurement counter is greater than the internal maximum limit;
HighOUT – the power meter is greater than the outer maximum limit.
Based on the results of collecting statistics, the following parameters are calculated:
NL is the result of the minimum power statistics.
NH is the result of the maximum power statistics.
PSP is the value of the stabilization point of the received power.
The SP control algorithm is shown in Fig. 6.
After collecting statistics, an analysis of the statistical pattern is carried out, which consists in comparing the sums of the numbers of NL and NH events, determined in accordance with the expressions:
NL = LowOUT · k + LowIN,
NH = HighOUT · k + HighIN,
where k is the coefficient of increase of peripheral sums (spaced by 3 dB from the boundaries of the dynamic range of the receiver).
In the case of NL > NH, the SP value increases, with the opposite ratio, it decreases. Thus, in the course of the proposed algorithm, the SP tends to a value at which, for the existing conditions (weather, distance), reaching the lower and upper boundaries of the receiver’s dynamic range is equally probable.
The coefficient k is necessary to increase the sensitivity of the algorithm to the most critical (closer Irec values to the border of the receiver dynamic range) events. For a reasonable choice of the value of k, the dependences of the correlation coefficient between the weight of the statistical pattern NL + NH and the previously obtained model estimate Perr for the lognormal distribution were calculated, over the entire field of values of SP and specified above.
The dependence of the correlation coefficient on the value of k is shown in Fig. 7.
Analysis of Fig. 7 shows that a fairly good (0.9) correlation of the pattern weight with Perr is achieved at k = 100. This value was used in the future when checking the performance of the algorithm.
To assess the performance of the SP tuning algorithm, “stress” measurements were carried out at a distance of 620 m with 20mm diaphragms installed on one of the terminals to ensure the level of instability of the receiving signal sufficient for fixing errors. On both terminals, in two series of measurements, the initial SP values were set, close to the upper (1,000 μW) and lower (10 μW) boundaries of the receiver’s dynamic range. Fig. 8 shows the dynamics of changes in the SP at two terminals of the 620 m line, as well as the level of losses (the number of lost packets per minute) in the communication channel.
It can be seen that the proposed algorithm for dynamic adjustment of the SP is fully operational and provides access to the optimal (in terms of minimizing losses in the communication channel) SP values.
The rather long (tens of minutes) time for tuning the SP is explained by the fact that the starting value of the SP was set as far as possible from the middle of the receiver’s dynamic range. When implementing the algorithm, the starting value of the SP was used, equal to the middle of the dynamic range of the receiver, which provides a much faster adjustment of the SP.
After determining the operability of the algorithm, a long-term (within three days) testing of its operation at a distance of 620 m was carried out. This testing showed a significant decrease in errors in comparison with the operation of a communication channel with a fixed value of the SP. So, if the error averaged 5 ∙ 10–10 per day, then when using the proposed algorithm, the error decreased to 9 ∙ 10–11.
CONCLUSIONS
A statistical analysis of the distribution of the values of the parameters of the wireless optical communication channel is carried out. He revealed a significant dependence of the channel parameters on the distance and operating conditions (state of the atmosphere). The analysis results also showed a discrepancy between the probability distribution of the received signal level and the known distributions.
A method is proposed for dynamically adjusting the power stabilization point at the receiver based on a statistical channel pattern that adequately describes the peripheral parts of the signal probability function at the receiver. Approbation of the algorithm has shown its effectiveness, which is expressed in a significant reduction in the level of errors in the communication channel. The use of the algorithm will improve the reliability of the transmitted data both when used in ground-based horizontal channels and in photonic systems of space instrumentation for information exchange on the lines “Spacecraft – Earth” [15]. Based on the results of the work, a patent was obtained for a method for adjusting the signal level at the receiver of an optical wireless communication terminal [16].
REFERENCES
Uysal M. Optical Wireless Communications. Springer. 2016. DOI 10.1007/978-3-319-30201-0
Djordjevic I. B. Advanced Optical and Wireless Communications Systems. Springer. 2017. DOI 10.1007/978-3-319-63151
First high capacity space-to-ground laser communications system for the new European external ISS Platform Bartolomeo [Internet]. [cited 12 Feb 2019]; URL: https://www.airbus.com/newsroom/press-releases/en/2018/03/first-high-capacity-space-to-ground-laser-communications-system-.html
Optical Payload for Lasercomm Science (OPALS). [Internet]. [cited 12 Feb 2019]; URL: https://www.nasa.gov/mission_pages/station/research/experiments/861.html
NASA Laser Communication Payload Undergoing Integration and Testing [Internet]. [cited 12 Feb 2019]; URL: https://www.nasa.gov/feature/goddard/2017/nasa-laser-communica-tion-payload-undergoing-integration-and-testing
Andrews L. C., Phillips R. L., Hopen C. Y. Laser Beam Scintillation with Applications. SPIE Press. 2001. DOI: 10.1117/3.412858
Majumdar A. K. Free-space laser communication performance in the atmospheric channel. J. Opt. Fiber. Commun. Rep. 2005; 2(4): 345–396. DOI: 10.1007/s10297-005-0054-0
10 Gbit/s ARTOLINK model [Internet]. [cited 12 Feb 2019]; URL: http://artolink.com/page/products/free_space_optics_Artolink_10Gbps.
Boev А. A., Kernosov М. Y., Kuznetsov S. N., Ognev B. I., Parshin А. А. The wireless communication channel with data rate 40 Gbps. Vestnik of Ryazan State Radio Engineering University. 62(4). 2017; 44–48. DOI: 10.21667/1995-4565-2017-62-4-44-48
Kuznetsov S. N., Ognev B. I., Poljakov S. Y., Yurko S. V. Wireless communications at 10Gb/s [Internet]. SPIE. Apr 2014. [cited 12 Feb 2019]; URL: http://spie.org/x108306.xml
Kuznetsov S. N., Polyakov S.YU., Oussama Alali, Bahaa Hashem. Besprovodnyj kanal 10 Gbit/s: klyuchevye osobennosti i rezul’taty testirovaniya. Sbornik trudov Mezhdunarodnoj nauchno-prakticheskoj konferencii “Innovacii v nauke, proizvodstve i obrazovanii”. – Ryazan. 2013; 83–93.
Arsenyan T. I., Zotov A. M., Korolenko P. V., Maganova M. S., Makarov V. G. Integral’nye raspredeleniya fluktuacij lazernogo izlucheniya v usloviyah peremezhaemosti atmosfernoj turbulentnosti. Atmospheric and Oceanic Optics. 2001; 14(8): 677–680.
Kernosov M. Y., Kuznetsov S. N., Ognev B. I., Polyakov S. Y., Shirobakin S. E. Analiz effektivnosti konstruktivnyh reshenij oborudovaniya vysokoskorostnoj opticheskoj besprovodnoj svyazi s uchetom vozdejstviya turbulentnoj atmosfery. Sbornik trudov XXV Mezhdunarodnoj Konferencii “Lazerno-informacionnye tekhnologii v medicine, biologii, geoekologii i transporte”. – Novorossiysk. 2017; 182–184.
Vorontsov M. A., Dudorov V. V., Zyryanova M. O. et al. Chastota poyavleniya oshibochnyh bitov v sistemah besprovodnoj opticheskoj svyazi s chastichno kogerentnym peredayushchim puchkom. Atmospheric and Oceanic Optics. 2012; 25(11): 936–940. DOI: 10.1134/S1024856013030159.
Ponomarev A. K., Romanov A. A., Tyulin A. E. Fotonnye tekhnologii v kosmicheskom priborostroenii. Raketno-kosmicheskoe priborostroenie i informacionnye sistemy. 2016; 3(2): 4–23. DOI: 10.17238/issn2409-0239.2016.2.4.
Patent RU2688664. Sposob regulirovki urovnya signala na priemnike terminala opticheskoj besprovodnoj svyazi /Parshin A. A., Kernosov M. Y., Kuznetsov S. N., Ognev B. I. URL: https://findpatent.ru/patent/268/2688664.html.
ABOUT AUTHORS
Kernosov M., info@moctkom.ru, JSС “Mostcom”, Ryazan, Russia.
Kuznetsov S., ksn@moctkom.ru, JSС “Mostcom”, Ryazan, Russia.
Parshin A., info@moctkom.ru, JSС “Mostcom”, Ryazan, Russia.
Ognev B., develop@moctkom.ru, JSС “Mostcom”, Ryazan, Russia.
Conflict of interest
The authors declare that they have no conflicts of interest.
Contribution of the team members
The article was prepared on the basis of many years of work by all members of the team of authors. Mathematical model, algorithm, software: Kuznetsov S. N., Parshin A. A. Layout preparation: Kernosov M. Yu., Ognev B. I. Experiments: Kernosov M. Yu., Kuznetsov S. N., Ognev B. I. Development and research was carried out at the expense of JSC Mostcom’s own funds.
M. Yu. Kernosov, S. N. Kuznetsov, B. I. Ognev, A. A. Parshin
Mostcom JSC, Ryazan, Russia
Analysis of the distribution probability of the received signal in a wireless optical communication line under various conditions shows a significant dependence of the error level in the communication channel on the level of power stabilization at the receiver. The significant difference between the statistics of the received signal level and the known distributions does not allow using them to approximate the probabilities of large deviations of the signal level from the mean value. An algorithm for evaluating the optimality and adjusting the level of stabilization of the receiving power based on the distribution statistics is proposed. According to experimental data, the proposed algorithm significantly reduces the level of errors in the atmospheric communication channel.
Keywords: laser communication terminal, free-space optics, atmospheric optical communication lines, wireless optical communications, power stabilization at an optical receiver, errors in a wireless communication channel
Received on: 10.07.2020
Accepted on: 24.08.2020
INTRODUCTION
The active development of wireless optical communications, which have certain advantages over other wireless solutions, in particular, is towards an increase in the data transmission rate [1, 2]. Optical wireless solutions have recently become especially relevant in high-speed communications of low-flying (LEO) spacecraft with the Earth [3–5]. The increase in the data transfer rate is unambiguously associated with a decrease in the dynamic range of the receiving devices. On the other hand, the dynamic range of the receiver directly determines the level of errors caused by atmospheric turbulence [6, 7]. In this regard, the optimization of the parameters of the transmit / receive channel becomes especially important.
In this paper, the study of optimization possibilities was carried out using Artolink wireless optical communication equipment designed to operate at speeds of 10 Gbps [8].
PROBLEM STATEMENT
One of the promising directions in the development of wireless optical communication systems is the creation of a wireless channel for transparent connection of high-speed interfaces of user equipment. This makes it easy to scale the wireless optical solution in the direction of increasing the transmission speed [9, 10]. On the other hand, a decrease in the dynamic range of the receiver, accompanying an increase in the transmission rate (mainly from the side of the lower boundary of the received powers – sensitivity), taking into account the influence of the atmosphere, requires the selection of the optimal value of the average signal level at the reception from the atmosphere, the stabilization point (SP). Typically, at 10 Gbps, the receiver (based on the PIN diode) has a dynamic range of about 21 dB (at a sensitivity of –18 dBm and a saturation power of 3 dBm). Taking into account the possible level of oscillations of the receiving signal in the atmosphere in a range comparable to the dynamic range of the receiver [11], the choice of the SP determines the level of digital errors in the channel. For a reasonable choice of the SP, it is necessary to know the statistics of the values of the received power (Iin) in order to minimize the number of events when Iin goes beyond the dynamic range of the receiver.
MODEL ASSESSMENT
At the first stage, a model assessment of the level of possible errors was carried out for various levels of signal instability and SP values. According to [12], the probability function of the receiving signal level when the laser radiation passes through the atmosphere is close to the lognormal law. Therefore, for a model assessment, the level of possible errors was calculated for a lognormal distribution.
The probability density of the lognormal distribution of the received power level relative to the mean value is described by the expression:
, (1)
where In is the received radiation power, normalized to the average value; σn is the root-mean-square deviation of the received radiation power, referred to its average value.
The value σn serving as a measure of the instability of the receiving signal is determined by the formula:
, (2)
where i is the ordinal number of the measurement in the sample; Iini is the value of the received power in the i-th dimension; Iavg the average value of the received power; N is the number of values in the sample.
The error level Perr was determined as the sum of the error probabilities due to a decrease in the signal level below the sensitivity level and an excess of the receiver saturation level according to the formula:
, (3)
where PSens is the probability of events with Irec below the receiver sensitivity level; PSat the probability of events with Irec above the saturation level of the receiver.
PSens and PSat were calculated using the following formulas:
(4)
(5)
where Imin is the receiver sensitivity level; Imax is the receiver saturation power level.
The calculations were performed for the σn levels at which Perr takes values from 10–16 to 10–2. The target range of Perr was reached at σn values from 0.1 to 0.5.
Taking into account that the middle of the receiver’s dynamic range corresponds to –6 dBm, the SP was set in the range from –10 to –2 dBm, corresponding to input signal levels from 100 to 600 μW. The dependences of Perr on SP for various values are shown in Fig. 1.
As seen from Fig. 1, the model dependence of the SP has an optimal value from the point of view of minimizing losses. The optimal SP value is in the region of –5 dBm, which differs from the midpoint of the receiver dynamic range of –6 dBm. The Perr level for optimal SP values is highly dependent on the level of signal instability, which is mainly determined by the distance and weather conditions on the track.
METHOD AND CONDITIONS OF MEASUREMENT
For measurements, a set of Artolink M1–10GE equipment was used, consisting of two terminals, each of which includes an optical unit (OU) – a transceiver, an interface unit (IU) and a connecting cable.
The software of the equipment allows to change the SP value, which is automatically maintained at the receiver of each terminal by adjusting the output power of the optical amplifier (OA) of the opposite transceiver using the service channel between the terminals. The Iin values with a frequency of 100 Hz were recorded on a computer connected to the interface for monitoring the state of the equipment on the IU.
The optoelectronic diagram of the transmitting and receiving path is shown in Fig. 2.
According to [13], the main factors causing fluctuations in the signal level at the reception are atmospheric turbulence and aperture limitation of the beam in the receiving plane. To analyze the statistics of Irec, the following series of measurements were carried out:
at a distance of 620 m at night and in the daytime;
at a distance of 620 m at night and in the daytime with aperture limitation (20 mm diaphragms – due to radiation divergence, this corresponds to fully opened apertures at a distance of 3,000 m) at one terminal corresponding to a distance of about 3,000 m;
at a distance of 1,600 m in the daytime;
at a distance of 2,800 m in the daytime.
In order to avoid the influence of changes in weather conditions during the measurements, each session was recorded for 10 minutes, which made it possible to record about 130,000 Iin values.
It should be noted that during all measurement sessions the following subsystems of the equipment worked: autotracking; adjustment of the direction of communication to the maximum of the received signal on each transceiver; stabilization of the received signal level by adjusting the power of the optical amplifier. Thus, the measurements were carried out under the conditions of fully functional operation of the equipment, which, in fact, was necessary, since the purpose of the work was to analyze and optimize the working version of the equipment.
MEASUREMENT RESULTS
“Fig. 3 shows the Iin values dynamics in μW for a SP of 350 μW at distances of 620, 1,600 and 2,800 m. As it can be seen, the depth of the falls of the values of Iin relative to the SP grows with an increase in the distance and reaches 10 dB at 2,800 m towards increasing its values.
To determine the actual dependence of the error level in the transmission channel on the SP value at a distance of 2,800 m, measurements were made of the dependence of the bit error level (BER) in the communication channel on the SP value. The measurement results are shown in Fig. 4.
It is clear that the SP clearly has an optimal value in terms of minimizing losses. In addition, a change in SP by only 50 μW leads to a change in BER level by an order of magnitude, which indicates the relevance of the SP optimization problem.
The choice of the optimal SP value depends on the nature of the distribution statistics of Iin values, the form of which, according to [14], significantly affects the BER level.
To quantify the nature of the statistics of the distribution of In values for all samples of In values obtained in the series of measurements, the values of the normalized root-mean square deviation σn and the asymmetry coefficient (Ca) of the distribution of the sample were calculated. The calculation of the σn value was carried out according to the formula (2). The following formula was used to calculate Ca:
. (6)
The results of calculating the parameters of statistics for various conditions are shown in the table. Their results show the following:
the instability of the receiving signal significantly depends on the turbulence of the atmosphere, increasing in the morning and decreasing at night;
σn increases with increasing distance, which, as shown in [9], is mainly determined by an increase in the aperture limitation of the receiving beam;
Ca in all series of measurements is nonzero and sometimes changes sign when going from a linear scale Iin to a logarithmic one.
Nonzero values of Ca and a change in sign indicate an intermediate character of the distribution density function In. Thus, with a small effect of atmospheric turbulence and aperture limitation, the distribution is closer to normal, and to lognormal with an increase in these effects.
Thus, the measurement results showed that a qualitative approximation of the real probability distribution by known types of distributions is not possible. In addition, one should take into account the fact that the integrals of the peripheral parts of the probability function play the main role in estimating the errors of the communication channel. In this regard, to solve the problem of optimization of the SP, a methods for assessing the peripheral parts of the distribution was required, which in the equipment should work in real time.
A solution was proposed that makes it possible to evaluate the peripheral parts of the specified function, since the error level is completely determined by the fraction of events of Irec going beyond the receiver’s dynamic range. Thus, SP tuning should minimize the number of these cases. Due to the fact that the measurement frequency of Iin in the equipment is 100 Hz, at an error level of 10–9, it will take 107 s to ensure that the dynamic range is exceeded, which corresponds to a period of about 2,800 hours.
To obtain a reasonable tuning period for the SP, it is sufficient to determine the statistics of achieving Iin at a certain distance from the boundaries of the dynamic range. However, due to their remoteness from the boundaries of the dynamic range, to take into account the shape of the “tails” of the probability function (PF), it is necessary to take into account their shape. For this, a method was proposed for assessing the peripheral parts of the PF based on the “statistical pattern” of the channel – an experimentally determined statistical image of the edges of the probability distribution Iin. To calculate the pattern, 4 parameters were used. They represent fractions of measurements with Iin values lying below and above certain boundaries, respectively (located at a certain distance from the boundaries of the receiver’s dynamic range). Fig. 5 shows an example of the experimental distribution of the number of measurements N depending on Iin, as well as the boundaries for calculating the Iin values to determine the statistical pattern of the channel.
As the lower boundaries of LowOUT and LowIN, the Pin values were chosen, which exceeded the receiver sensitivity by 3 and 6 dB (2 and 4 times), respectively. As the upper boundaries of HighOUT and HighIN – Iin, lower saturation powers of the receiver by 3 and 6 dB, respectively.
To test the proposed method for adjusting the SP, an algorithm for changing the SP was implemented depending on the result of the analysis of the statistical pattern obtained in 1 minute of equipment operation.
Statistics are collected for N = 30,000 receive power measurements using the following counters:
LowIN – power meter is less than the inner minimum limit;
LowOUT – power meter is less than the outer minimum limit;
HighIN – power measurement counter is greater than the internal maximum limit;
HighOUT – the power meter is greater than the outer maximum limit.
Based on the results of collecting statistics, the following parameters are calculated:
NL is the result of the minimum power statistics.
NH is the result of the maximum power statistics.
PSP is the value of the stabilization point of the received power.
The SP control algorithm is shown in Fig. 6.
After collecting statistics, an analysis of the statistical pattern is carried out, which consists in comparing the sums of the numbers of NL and NH events, determined in accordance with the expressions:
NL = LowOUT · k + LowIN,
NH = HighOUT · k + HighIN,
where k is the coefficient of increase of peripheral sums (spaced by 3 dB from the boundaries of the dynamic range of the receiver).
In the case of NL > NH, the SP value increases, with the opposite ratio, it decreases. Thus, in the course of the proposed algorithm, the SP tends to a value at which, for the existing conditions (weather, distance), reaching the lower and upper boundaries of the receiver’s dynamic range is equally probable.
The coefficient k is necessary to increase the sensitivity of the algorithm to the most critical (closer Irec values to the border of the receiver dynamic range) events. For a reasonable choice of the value of k, the dependences of the correlation coefficient between the weight of the statistical pattern NL + NH and the previously obtained model estimate Perr for the lognormal distribution were calculated, over the entire field of values of SP and specified above.
The dependence of the correlation coefficient on the value of k is shown in Fig. 7.
Analysis of Fig. 7 shows that a fairly good (0.9) correlation of the pattern weight with Perr is achieved at k = 100. This value was used in the future when checking the performance of the algorithm.
To assess the performance of the SP tuning algorithm, “stress” measurements were carried out at a distance of 620 m with 20mm diaphragms installed on one of the terminals to ensure the level of instability of the receiving signal sufficient for fixing errors. On both terminals, in two series of measurements, the initial SP values were set, close to the upper (1,000 μW) and lower (10 μW) boundaries of the receiver’s dynamic range. Fig. 8 shows the dynamics of changes in the SP at two terminals of the 620 m line, as well as the level of losses (the number of lost packets per minute) in the communication channel.
It can be seen that the proposed algorithm for dynamic adjustment of the SP is fully operational and provides access to the optimal (in terms of minimizing losses in the communication channel) SP values.
The rather long (tens of minutes) time for tuning the SP is explained by the fact that the starting value of the SP was set as far as possible from the middle of the receiver’s dynamic range. When implementing the algorithm, the starting value of the SP was used, equal to the middle of the dynamic range of the receiver, which provides a much faster adjustment of the SP.
After determining the operability of the algorithm, a long-term (within three days) testing of its operation at a distance of 620 m was carried out. This testing showed a significant decrease in errors in comparison with the operation of a communication channel with a fixed value of the SP. So, if the error averaged 5 ∙ 10–10 per day, then when using the proposed algorithm, the error decreased to 9 ∙ 10–11.
CONCLUSIONS
A statistical analysis of the distribution of the values of the parameters of the wireless optical communication channel is carried out. He revealed a significant dependence of the channel parameters on the distance and operating conditions (state of the atmosphere). The analysis results also showed a discrepancy between the probability distribution of the received signal level and the known distributions.
A method is proposed for dynamically adjusting the power stabilization point at the receiver based on a statistical channel pattern that adequately describes the peripheral parts of the signal probability function at the receiver. Approbation of the algorithm has shown its effectiveness, which is expressed in a significant reduction in the level of errors in the communication channel. The use of the algorithm will improve the reliability of the transmitted data both when used in ground-based horizontal channels and in photonic systems of space instrumentation for information exchange on the lines “Spacecraft – Earth” [15]. Based on the results of the work, a patent was obtained for a method for adjusting the signal level at the receiver of an optical wireless communication terminal [16].
REFERENCES
Uysal M. Optical Wireless Communications. Springer. 2016. DOI 10.1007/978-3-319-30201-0
Djordjevic I. B. Advanced Optical and Wireless Communications Systems. Springer. 2017. DOI 10.1007/978-3-319-63151
First high capacity space-to-ground laser communications system for the new European external ISS Platform Bartolomeo [Internet]. [cited 12 Feb 2019]; URL: https://www.airbus.com/newsroom/press-releases/en/2018/03/first-high-capacity-space-to-ground-laser-communications-system-.html
Optical Payload for Lasercomm Science (OPALS). [Internet]. [cited 12 Feb 2019]; URL: https://www.nasa.gov/mission_pages/station/research/experiments/861.html
NASA Laser Communication Payload Undergoing Integration and Testing [Internet]. [cited 12 Feb 2019]; URL: https://www.nasa.gov/feature/goddard/2017/nasa-laser-communica-tion-payload-undergoing-integration-and-testing
Andrews L. C., Phillips R. L., Hopen C. Y. Laser Beam Scintillation with Applications. SPIE Press. 2001. DOI: 10.1117/3.412858
Majumdar A. K. Free-space laser communication performance in the atmospheric channel. J. Opt. Fiber. Commun. Rep. 2005; 2(4): 345–396. DOI: 10.1007/s10297-005-0054-0
10 Gbit/s ARTOLINK model [Internet]. [cited 12 Feb 2019]; URL: http://artolink.com/page/products/free_space_optics_Artolink_10Gbps.
Boev А. A., Kernosov М. Y., Kuznetsov S. N., Ognev B. I., Parshin А. А. The wireless communication channel with data rate 40 Gbps. Vestnik of Ryazan State Radio Engineering University. 62(4). 2017; 44–48. DOI: 10.21667/1995-4565-2017-62-4-44-48
Kuznetsov S. N., Ognev B. I., Poljakov S. Y., Yurko S. V. Wireless communications at 10Gb/s [Internet]. SPIE. Apr 2014. [cited 12 Feb 2019]; URL: http://spie.org/x108306.xml
Kuznetsov S. N., Polyakov S.YU., Oussama Alali, Bahaa Hashem. Besprovodnyj kanal 10 Gbit/s: klyuchevye osobennosti i rezul’taty testirovaniya. Sbornik trudov Mezhdunarodnoj nauchno-prakticheskoj konferencii “Innovacii v nauke, proizvodstve i obrazovanii”. – Ryazan. 2013; 83–93.
Arsenyan T. I., Zotov A. M., Korolenko P. V., Maganova M. S., Makarov V. G. Integral’nye raspredeleniya fluktuacij lazernogo izlucheniya v usloviyah peremezhaemosti atmosfernoj turbulentnosti. Atmospheric and Oceanic Optics. 2001; 14(8): 677–680.
Kernosov M. Y., Kuznetsov S. N., Ognev B. I., Polyakov S. Y., Shirobakin S. E. Analiz effektivnosti konstruktivnyh reshenij oborudovaniya vysokoskorostnoj opticheskoj besprovodnoj svyazi s uchetom vozdejstviya turbulentnoj atmosfery. Sbornik trudov XXV Mezhdunarodnoj Konferencii “Lazerno-informacionnye tekhnologii v medicine, biologii, geoekologii i transporte”. – Novorossiysk. 2017; 182–184.
Vorontsov M. A., Dudorov V. V., Zyryanova M. O. et al. Chastota poyavleniya oshibochnyh bitov v sistemah besprovodnoj opticheskoj svyazi s chastichno kogerentnym peredayushchim puchkom. Atmospheric and Oceanic Optics. 2012; 25(11): 936–940. DOI: 10.1134/S1024856013030159.
Ponomarev A. K., Romanov A. A., Tyulin A. E. Fotonnye tekhnologii v kosmicheskom priborostroenii. Raketno-kosmicheskoe priborostroenie i informacionnye sistemy. 2016; 3(2): 4–23. DOI: 10.17238/issn2409-0239.2016.2.4.
Patent RU2688664. Sposob regulirovki urovnya signala na priemnike terminala opticheskoj besprovodnoj svyazi /Parshin A. A., Kernosov M. Y., Kuznetsov S. N., Ognev B. I. URL: https://findpatent.ru/patent/268/2688664.html.
ABOUT AUTHORS
Kernosov M., info@moctkom.ru, JSС “Mostcom”, Ryazan, Russia.
Kuznetsov S., ksn@moctkom.ru, JSС “Mostcom”, Ryazan, Russia.
Parshin A., info@moctkom.ru, JSС “Mostcom”, Ryazan, Russia.
Ognev B., develop@moctkom.ru, JSС “Mostcom”, Ryazan, Russia.
Conflict of interest
The authors declare that they have no conflicts of interest.
Contribution of the team members
The article was prepared on the basis of many years of work by all members of the team of authors. Mathematical model, algorithm, software: Kuznetsov S. N., Parshin A. A. Layout preparation: Kernosov M. Yu., Ognev B. I. Experiments: Kernosov M. Yu., Kuznetsov S. N., Ognev B. I. Development and research was carried out at the expense of JSC Mostcom’s own funds.
Readers feedback
