Optimising models for prediction of tropospheric scintillation on satellite links
2020; Institution of Engineering and Technology; Volume: 56; Issue: 11 Linguagem: Inglês
10.1049/el.2020.0216
ISSN1350-911X
AutoresNelson A. Pérez-García, Francklin Rivas, Leidy Marian Rujano, Ángel Pinto, José Manuel Torres Torres, José Aguilar, Eduardo José Ramírez, Jaime Vélez-Zapata,
Tópico(s)Atmospheric aerosols and clouds
ResumoElectronics LettersVolume 56, Issue 11 p. 577-579 Wireless communicationsFree Access Optimising models for prediction of tropospheric scintillation on satellite links N.A. Pérez-García, Corresponding Author N.A. Pérez-García perezn@ula.ve Grupo de Investigación de Telecomunicaciones (GITEL) and Postgrado en Telecomunicaciones, Facultad de Ingeniería, Universidad de Los Andes, Sector La Hechicera, 5101, Mérida, VenezuelaSearch for more papers by this authorF.R. Rivas, F.R. Rivas Grupo de Investigación de Telecomunicaciones (GITEL) and Postgrado en Telecomunicaciones, Facultad de Ingeniería, Universidad de Los Andes, Sector La Hechicera, 5101, Mérida, VenezuelaSearch for more papers by this authorL.M. Rujano, L.M. Rujano Grupo de Investigación de Telecomunicaciones (GITEL) and Postgrado en Telecomunicaciones, Facultad de Ingeniería, Universidad de Los Andes, Sector La Hechicera, 5101, Mérida, VenezuelaSearch for more papers by this authorA.D. Pinto, A.D. Pinto Escuela de Ingeniería de Sistemas, Universidad del Sinú, Carrera 1w No. 38-153, Barrio Juan XIII, Montería, 230001 ColombiaSearch for more papers by this authorJ.M. Torres, J.M. Torres Escuela de Ingeniería de Sistemas, Universidad del Sinú, Carrera 1w No. 38-153, Barrio Juan XIII, Montería, 230001 ColombiaSearch for more papers by this authorJ. Aguilar-Castro, J. Aguilar-Castro Centro de Estudios en Microelectrónica y Sistemas Distribuidos (CEMISID), Facultad de Ingeniería, Universidad de Los Andes, Sector La Hechicera, 5101, Mérida, Venezuela J. Aguilar-Castro: Also with Grupo de Investigación, Desarrollo e Innovación en Tecnologías de la Información y las Comunicaciones (GIDITIC), Universidad EAFIT, Carrera 49, No. 7, Sur-50, 050022, Medellín, ColombiaSearch for more papers by this authorE.J. Ramírez, E.J. Ramírez Grupo de Investigación de Telecomunicaciones (GITEL) and Postgrado en Telecomunicaciones, Facultad de Ingeniería, Universidad de Los Andes, Sector La Hechicera, 5101, Mérida, VenezuelaSearch for more papers by this authorJ. Vélez-Zapata, J. Vélez-Zapata Departamento de Ciencias de la Computación y Electrónica, Universidad de la Costa, Calle 58, No, 55-66, Barranquilla, 080002 ColombiaSearch for more papers by this author N.A. Pérez-García, Corresponding Author N.A. Pérez-García perezn@ula.ve Grupo de Investigación de Telecomunicaciones (GITEL) and Postgrado en Telecomunicaciones, Facultad de Ingeniería, Universidad de Los Andes, Sector La Hechicera, 5101, Mérida, VenezuelaSearch for more papers by this authorF.R. Rivas, F.R. Rivas Grupo de Investigación de Telecomunicaciones (GITEL) and Postgrado en Telecomunicaciones, Facultad de Ingeniería, Universidad de Los Andes, Sector La Hechicera, 5101, Mérida, VenezuelaSearch for more papers by this authorL.M. Rujano, L.M. Rujano Grupo de Investigación de Telecomunicaciones (GITEL) and Postgrado en Telecomunicaciones, Facultad de Ingeniería, Universidad de Los Andes, Sector La Hechicera, 5101, Mérida, VenezuelaSearch for more papers by this authorA.D. Pinto, A.D. Pinto Escuela de Ingeniería de Sistemas, Universidad del Sinú, Carrera 1w No. 38-153, Barrio Juan XIII, Montería, 230001 ColombiaSearch for more papers by this authorJ.M. Torres, J.M. Torres Escuela de Ingeniería de Sistemas, Universidad del Sinú, Carrera 1w No. 38-153, Barrio Juan XIII, Montería, 230001 ColombiaSearch for more papers by this authorJ. Aguilar-Castro, J. Aguilar-Castro Centro de Estudios en Microelectrónica y Sistemas Distribuidos (CEMISID), Facultad de Ingeniería, Universidad de Los Andes, Sector La Hechicera, 5101, Mérida, Venezuela J. Aguilar-Castro: Also with Grupo de Investigación, Desarrollo e Innovación en Tecnologías de la Información y las Comunicaciones (GIDITIC), Universidad EAFIT, Carrera 49, No. 7, Sur-50, 050022, Medellín, ColombiaSearch for more papers by this authorE.J. Ramírez, E.J. Ramírez Grupo de Investigación de Telecomunicaciones (GITEL) and Postgrado en Telecomunicaciones, Facultad de Ingeniería, Universidad de Los Andes, Sector La Hechicera, 5101, Mérida, VenezuelaSearch for more papers by this authorJ. Vélez-Zapata, J. Vélez-Zapata Departamento de Ciencias de la Computación y Electrónica, Universidad de la Costa, Calle 58, No, 55-66, Barranquilla, 080002 ColombiaSearch for more papers by this author First published: 01 May 2020 https://doi.org/10.1049/el.2020.0216 AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinkedInRedditWechat Abstract A phenomenon that also causes impairment in the received signal strength of the RF signal in satellite links operating at frequencies above 10 GHz, especially in those systems that operate at higher frequencies with small fade margins, is the tropospheric scintillation that occurs in the lower layer of the troposphere. In order to estimate the intensity, i.e. the variance between the scintillation amplitude fluctuations, there are several models in the literature, whose accuracy depends on the locality in which the models are implemented. In this Letter, new models are developed for the prediction of tropospheric scintillation that adjusts to localities in Spain, specifically Madrid city, based on measurements of the aforementioned phenomenon reported from Spain and the bios-inspired technique Cuckoo Search (CS). The results obtained, evaluated in terms of the root mean square error, were totally satisfactory, being the most outstanding cases the improved versions of the Ortgies-T, Statistical Temperature and Humidity 2 and Statistical Temperature and Refractivity 2 models. Introduction Currently, the use of higher frequency bands for satellite communications, such as the Ka, Q, V and W bands, have high demand, due to spectrum congestion on lower frequencies and to the advent of technologies (for example, high-throughput satellite systems), which require larger bandwidths [1]. In these bands, the performance of the satellite communication systems depends basically on the propagation characteristics of the atmosphere, being one of the effects to consider the tropospheric scintillation. Although, as well known, the rain has the highest impact on these systems operating at frequencies above 10 GHz, the effect of tropospheric scintillation is significant for systems planned and dimensioned at high frequencies with a low fade margin [2-4]. Considered to be a short timescale effect, similar to that produced by the multipath phenomenon in wireless systems operating at frequencies 80%. Among the results obtained by [14] that are of interest for the development of the new tropospheric scintillation prediction model, are those related to the variance, m2, of the aforementioned scintillation. Also, will be used the monthly results corresponding to temperature, T, relative humidity, U, and atmospheric pressure, P, parameters estimated in the same work [14]. Predicting tropospheric scintillation models In the prediction of the monthly mean value of tropospheric scintillation intensity, m, by Karasawa, Otung-1, Otung-2, Rahim and ITU models, the general form is [6, 9, 10, 12, 13] (1)where a and b are regression constants whose values depend on the model, Nwet is the wet term (can be obtained directly from the Recommendation ITU-R P.453–14 [18], or calculated using T and U in Karasawa model and Rahim model or T, U and P in the other models), f is the system operating frequency, ε means the elevation angle of the earth station antenna, the values of the α and β parameters are particular to each model (fixed values in all models considered, except in the Otung-2, in which both parameters depend on the angle ε), and is G(x) or g(x), with x depending on the diameter of the receiving antenna (D), the frequency (f), the efficiency of the receiving antenna (η), and the inclined distance from the receiver to the height of the turbulence layer (L). In the Ortgies-N, Ortgies-T, STH2 and STN2 models, m is of the form [8, 11] (2)where ρ is function of Nwet in Ortgies-N model, T in Ortgies-T model, U and T in STH2 model and Nwet and T in STN2 model. General concepts of CS The CS optimisation algorithm is based on the parasitic and forced reproduction mechanism of some species of cuckoo birds, which lay their eggs in the nests of other birds (often other species), even expelling the eggs from the invaded nest to increase the probability that theirs will be born. If the host bird becomes aware that the egg is not its own, it will either kick the foreign egg out or abandon its nest and establish a new one, elsewhere [15-17]. The CS algorithm must comply with the following three basic rules [15]: (i) a global search process (known as diversification or exploration), in which each cuckoo bird will lay only one egg at a time, and choose the nest randomly, (ii) a local search process (known as intensification or exploitation), in which the best nests, i.e. the best solutions, will advance to the next generation and (iii) the number of host nests is fixed, and an egg laid in a nest can be discovered by the host bird with an abandon probability, pa∈ [0,1]. Here, the host bird chooses between either abandon the nest or throw the foreign egg away. The local search in the CS algorithm is implemented as follows [17]: (3)where and represent the ith nest location in the kth current generation and the (kth + 1) new generation, respectively, α > 0 means the scaling factor, s is the step length, the product refers to the point to point multiplications, H(u) represents the Heaviside function, r1 is a random number generated using uniform distribution, is a random selected solution and represents the current global best nest. In (2), it is important to emphasise that, if pa < r1, then , and, consequently, . For the global search, in the CS algorithm is used Lévy flight [15, 16] (4)where β is the distribution factor. In general, a Lévy flight is a random walk in space, in which the step length (s) follows a Lévy distribution. In most problems, the Mantegna algorithm is used to define s. This algorithm is based on samples, u and v, drawn from a normal distribution and , respectively, with a mean equal to zero for both and standard deviation and [16]. New models and comparative test To tune the tropospheric scintillation models mentioned above, variants are incorporated in them in order to minimise the cost function, i.e. the RMSE values [19] specifically between the estimated variance by the models and the measured variance values. The variants with which the best adjustments were obtained for each model were (5)for Karasawa, Otung-1, Otung-2, Rahim and ITU models. (6)for Ortgies-N, Ortgies-T, STH2 and STN2 models, where (7) (8) (9) (10)Also, the pair (x; y) in (5)–(10) are the adjustment parameters for each model, to be obtained using the CS algorithm. Thus, with a number of parameters to be found in each of the new models equal to 2, the best fit for each case was obtained for n = 10 nests, 200 iterations, α = 0.01, pa = 0.25, stop criterion equal to 1 × 10−8 and β = 1.5. The values obtained for the adjustment coefficients of each modified model are shown in Table 1. Table 1. Adjustment parameter values Model Kar ON OT Otu-1 Otu-2 STH2 STN2 Rahim ITU x 5.345 0.520 1.427 3.979 5.892 0.984 1.215 5.487 4.643 y 0.271 0.762 0.748 0.501 0.725 1.224 1.263 0.231 0.360 Kar = Karasawa; ON = Ortgies-N; OT = Ortgies-T; Otu-1 = Otung-1; Otu-2 = Otung-2. Figs. 1 and 2 show the performance of the original and modified versions of the models, respectively. Before optimisation, in the fall and winter seasons in Spain (from October to March), all models, except Otung-1, Otung-2 and ITU models, underestimate the estimated tropospheric scintillation variance values, with better performance for Rahim and Karasawa models; while for the months of spring and summer, i.e. from April to September, the Otung-2 model shows the best performance, although with a tendency to overestimate the measured values, while the other models underestimate the aforementioned values with an evident poor performance. After the optimisation, except for the Rahim, Karasawa and Otung-1 models, the remaining models improved their performance in the fall and winter seasons, while in the spring and summer seasons all models improved their performance, with an excellent approximation to the values measured by Ortgies-T-I, STH2-I and STN2-I models. Fig 1Open in figure viewerPowerPoint Measurements and estimated mean monthly values of the tropospheric scintillation variance before the optimisation Fig 2Open in figure viewerPowerPoint Measurements and estimated mean monthly values of the tropospheric scintillation variance after the optimisation The previous assessments are confirmed by the RMSE values, between the measured values and the predicted values by the models, summarised in Table 2, which also shows that considering all the annual periods of the measurements carried out, the RMSE value is improved for all the models after they were modified, with the smallest error (<1 × 10−3) and an excellent accuracy by Ortgies-T, STH2 and STN2 modified models. Table 2. RMSE values before and after the optimisation (× 103 dB2) Model Kar ON OT Otu-1 Otu-2 STH2 STN2 Rahim ITU all annual measurement periods before 2.820 4.163 3.481 2.059 2.124 2.903 2.825 2.692 1.838 after 1.642 1.758 0.939 1.691 1.235 0.880 0.821 1.636 1.686 fall and winter months before 0.611 1.799 1.893 1.003 2.494 1.450 1.496 0.504 1.867 after 1.715 1.627 0.659 1.748 1.303 0.839 0.715 1.657 1.747 spring and summer periods before 3.966 5.646 4.578 2.754 1.709 3.869 3.732 3.797 1.832 after 1.587 1.900 1.162 1.655 1.180 0.929 0.924 1.634 1.645 Conclusion In this Letter, new models for the estimation of tropospheric scintillation in satellite links were obtained, tuning existing models to measurements conducted in Spain and the CS nature-inspired technique. The results show an improvement, in terms of the RMSE values, for all the new models for the total annual period of measurements, especially for the Ortgies-T-I, STH2-I and STN-I models, in which the introduced variants that provided the best results include the direct adjustment of the temperature (T). Lastly, for the spring and summer season, all models also improved their performance, while for the fall and winter months there were improvements also, except for the Karasawa, Otung-1 and Rahim models. References 1Chong, C.K., Layman, D.A., McGeary, W.L. et. al.,: 'Q/V-band high-power uplink helix TWT for future high-data-rate communications', IEEE Trans. Electron. Devices, 2018, 65, (6), pp. 2201– 2205 (https://doi/org/10.1109/TED.2018.2818619) 2Ippolito, L.J.: ' Satellite communications systems engineering. Atmospheric effects, satellite link design and system performance' ( John Wiley & Sons, Chichester, UK, 2017, 2nd edn.) 3Omotosho, T.V., Akinwumi, S.A, Usikalu, M.R et. al.,: 'Analysis of non-rainy attenuation on earth-space path in Ota, Southwest Nigeria', J. Phys. Conf. Ser., 2017, 852, (012039), pp. 1– 6 4Madhuri, A.S., Immadi, G., Narayana, M.V.: 'Estimation of cumulative distribution of scintillation effect on Ku band frequencies for one of the tropical regions using various models', J. Eng. Sci. Tech. Rev., 2018, 11, (1), pp. 151– 155 (https://doi/org/10.25103/jestr.111.18) 5Chen, C.Y., Singh, M.J.: 'Comparison of tropospheric scintillation prediction models of the Indonesian climate', Earth Planets Space, 2014, 66, (64), pp. 1– 12 6Karasawa, Y., Yamada, M., Allnut, J.E.: 'A new prediction method for tropospheric scintillation on earth-space paths', IEEE Trans. Antennas Propag., 1988, 36, (11), pp. 1608– 1614 (https://doi/org/10.1109/8.9712) 7Rivas-Davila, F.R., Rujano-Molina, L.M., Perez-Garcia, N.A. et. al.,: 'Predicción del centelleo troposférico producido por el satélite Simón Bolívar en Mucuchíes, Venezuela, en la banda Ka', Ingeniería Al Día, 2018, 4, (2), pp. 49– 71 8Ortgies, G.: 'Frequency dependence of slant-path amplitude scintillations', Electron. Lett., 1993, 29, (25), pp. 2219– 2220 (https://doi/org/10.1049/el:19931490) 9Otung, I.E.: 'Prediction of tropospheric amplitude scintillation on a satellite link', IEEE Trans. Antennas Propag., 1996, 44, (12), pp. 1600– 1608 (https://doi/org/10.1109/8.546246) 10Otung, I.E., Savvaris, A.: 'Estimating tropospheric scintillation intensity on earth-space propagation paths', Electron. Lett., 2006, 42, (7), pp. 381– 382 (https://doi/org/10.1049/el:20064281) 11Marzano, F.S., Carlo, R., Alessio, B. et. al.,: 'Assessment of model-based scintillation variance prediction on long-term basis using Italsat satellite measurements', Int. J. Satell. Commun., 1999, 17, (1), pp. 17– 36 (doi: https://doi/org/10.1002/(SICI)1099-1247(199901/02)17:1 3.0.CO;2-9) 12Rahim, N.B.A., Islam, M.R., Mandeep, J.S. et. al.,: 'Tropospheric scintillation prediction models for a high elevation angle based on measured data from a tropical region', J. Atmos. Sol. Terr. Phys., 2013, 105–106, pp. 91– 96 (https://doi/org/10.1016/j.jastp.2013.08.005) 13 Recommendation ITUR P.618–13: ' Propagation data and prediction methods required for the design of earth-space telecommunication systems' (International Telecommunications Union, Geneva, Switzerland, 2017) 14Garcia-Merchan, E.: ' Caracterización experimental del centelleo troposférico en banda Ka', Final Degree Project, Universidad Politécnica de Madrid, 2014 15Feng, Y., Zhou, J., Mo, L. et. al.,: 'A gradient-based cuckoo search algorithm for a reservoir-generation scheduling problem', Algorithms, 2018, 11, (36), pp. 1– 21 16Ding, J., Wang, Q., Zhang, Q. et. al.,: 'A hybrid particle swarm optimization-cuckoo search algorithm and its engineering applications', Math. Probl. Eng., 2019, 2019, (ID 5213759), pp. 1– 12 17Yang, X.S., He, X.S.: ' Mathematical foundations of nature-inspired algorithms' ( Springer, Cham, Switzerland, 2019) 18 Recommendation ITUR P. 453–14: ' The radio refractive index: its formula and refractivity data', (International Telecommunications Union, Geneva, Switzlerand, 2019) 19Torres-Tovio, J.M., Pérez-García, N.A., Pinto-Mangones, A.D. et. al.,: ' Novel Lee model for prediction of propagation path loss in digital terrestrial television systems in Montevideo city, Uruguay', The International Conference on Advances in Emerging Trends and Technologies (ICAETT 2019), Guayaquil, Ecuador, March 2019, pp. 542– 553 Volume56, Issue11May 2020Pages 577-579 FiguresReferencesRelatedInformation
Referência(s)