Microsoft Word - 24-2630_s Engineering, Technology & Applied Science Research Vol. 9, No. 2, 2019, 3998-4001 3998 www.etasr.com Eli-Chukwu & Onoh: Experimental Study on the Impact of Weather Conditions on Wide Code … Experimental Study on the Impact of Weather Conditions on Wide Code Division Multiple Access Signals in Nigeria Ngozi Clara Eli-Chukwu Department of Electrical & Electronics Engineering, Alex Ekwueme Federal University, Ndufu-Alike, Ebonyi, Nigeria ngozieli@gmail.com G. N. Onoh Electrical & Electronics Engineering, Enugu State University of Science and Technology, Enugu, Nigeria onohgn@gmail.com Abstract—In cellular network activities, before a site is integrated it is expected that each cell of the site meets the Nigerian Communication Commission (NCC) standard of ≥98% for both service accessibility and call completion rate which in turn depicts a ≤2% in both blocked call rate (BCR) and dropped call rate (DCR). It is suggested that weather conditions have a very strong negative effect on the performance of wideband code division multiple access (WCDMA) network as it could lead to signal attenuation or change the polarization. In this paper, we study the impact of weather conditions on WCDMA network in Nigeria. To achieve this, network samples (log-files) were collected weekly during a driving test in Enugu State Nigeria for a period of five years for both rainy and dry seasons, in which blocked and dropped calls were extracted. Results show that during adverse weather conditions, BCR and DCR rise greater than 8% and 4% respectively. Although with a slight relationship between the weather conditions, the weather condition during the dry season has a better-blocked call rate of 8.76% in comparison with the rainy season with 12.89%. Calls tend to drop more during the dry season. From the outcome of the experiment, a model was developed for predicting an unknown network call statistics variables. Keywords-blocked call; cellular; dropped call; dry season; rainy season; weather; WCDMA I. INTRODUCTION In cellular networks, trying for optimization is considered a constant. One of the major reasons for constant optimization is the weather. Often, weather conditions affect major key performance indicators (KPI) such as accessibility and retainability used by operators and subscribers to assess network performance. During various weather conditions, it is expected for the WCDMA cellular network to maintain its performance and meet the Nigerian Communication Commission (NCC) threshold of both BCR and DCR ≤2%. This paper studies the impacts of weather conditions on WCDMA network and compares them. GSM signal strength varies with respect to weather parameters. Tropospheric delays such as humidity, pressure and temperature affect the strength of the transmitted signal [1]. Distance estimation based on the received signal strength of a wireless radio is susceptible to radio propagation conditions, particularly during periods of precipitation [2]. Movement of mobile station (MS) affects the signal received from the base transceiver system (BTS). Decreased signal reception conditions, if the channel is exposed to rain, occur. Rain attenuation as color noise affects signal quality, mainly 3G network signal transmission, and can be grouped into drizzling rain, straight form, medium, heavy convective and storms, have the effect of their own in the process of signal transmission. [3]. Many wireless sensor networks operating outdoors are exposed to changing weather conditions, which may cause severe degradation in system performance. Therefore, it is essential to explore the factors affecting radio link quality in order to mitigate their impact [4]. There is a relationship between atmospheric conditions and speech quality [5]. Since the cellular GSM networks are one of the most commonly used communication technologies today, the quality of speech in these networks is a topic of great significance. Many advances and approaches have been introduced in the field of speech quality during the last decade, most of them focusing on IP networks, where speech quality is influenced by every single network node through which the communication passes. There is a bond between speech quality in GSM networks and weather conditions and a greater bond between rain density and speech quality [6]. The most classical approach of determining rain attenuation for radio-wave frequency has been to theoretically determine the specific attenuation. At frequency over 10GHz, rain and precipitation can influence the attenuation a lot. The effect of atmospheric attenuation between the source and destination over wireless communication is of major concern and proper site visit and control method are required so that the performance can be increased [7]. Rainfall is a natural phenomenon whose temporal and geographical distribution varies widely. Wireless communications suffer losses in network quality during rainfall which can affect the regional communication for a while. Growing concerns of climatic change also encourage the study of the effects of natural phenomena like rainfall on other measurable parameters Corresponding author: Ngozi Clara Eli-Chukwu Engineering, Technology & Applied Science Research Vol. 9, No. 2, 2019, 3998-4001 3999 www.etasr.com Eli-Chukwu & Onoh: Experimental Study on the Impact of Weather Conditions on Wide Code … [8]. The link availability of outdoor radio systems is often affected by atmospheric conditions such as rain and snow. The effects of rainfall on wireless transmissions are particularly noticeable at data rates. As the rate of rainfall increases, more disruption is caused to the outdoor link [9]. Some dependencies between weather conditions and receive level were studied in [10]. There is a relationship between refractivity from rainfall and propagation of the GSM radio signal as greater refractivity means lower signal quality and vice versa [11]. The GSM technology is the most widely utilized communication standard which it is now coming to its bandwidth limitations especially in big cities and densely populated areas. Under such circumstances, even a minor weather change could be a decisive factor causing changes in the quality of service [12]. Attenuation in tropical regions is underestimated by existing prediction methods based on experimental data from temperate climates [13]. Sunny and rainy weather conditions have first- order main effects on user equipment [14]. The significance here is that various forms of precipitation such as rain, snow, cloud and fog absorb and scatter electromagnetic energy leading to attenuation in its signal strength. Harmattan precipitation intensity may be so great that visibility at ground level is reduced to less than a hundred meters while inflicting attenuation significantly [15]. Certain combinations of the constituents in weather can cause radio signals to be heard hundreds of miles beyond the ordinary range of radio communications. Tropical weather has significant effects on radio signal where the highest correlation values for each factor are 0.70756 for solar radiation, 0.6285 for humidity, 0.4344 for wind speed, 0.3850 for rain rate, and 0.3339 for temperature [16]. Terrestrial and Earth-space links operating at bands higher than 10GHz inevitably suffer severe signal degradation due to rain fade, particularly in the tropics [17]. This study focuses on ascertaining and comparing the impact of various weather conditions on WCDMA network. II. EXPERIMENTAL SETUP A. Method The drive test method was used to characterize the network [18]. The experimental setup uses the Testing Equipment for Mobile System (TEMS) V13.0 software installed on a laptop, a TEMS mobile phone, a GPS and a power inverter. Voice calls were made for 120secs on the MTNN network by the mobile phone. The test covered the Enugu metropolis in Enugu State, Nigeria for a period of 5 years, covering both rainy and dry seasons. B. Network Characterization Parameter Although the experimental technique reports many parameters, the parameters of interest are BCR and DCR. BCR explains the rate at which the user equipment (UE) is unable to access the network when a call is attempted. It is expressed as: %��� = ���∗ �� � (1) where ��� is the number of blocked calls and ��� the number of successful calls. DCR is the rate at which established calls ends abruptly without the knowledge of both call originator and terminator: %��� = ���∗ �� �� (2) where ��� is the number of is dropped calls and ��� is the number of successful calls. C. Network Characterization Results Table I shows the call statistics and radio environment results for the five year period. TABLE I. CALL STATISTICS KPI RESULT FROM NETWORK CHARACTERIZATION. Year Weeks Rainy season Dry/Harmattan season BCR DCR BCR DCR 2015 Wk1 7.14 3.66 7.43 2.06 Wk2 3.19 3.12 2.1 2.27 Wk3 1.01 2.6 1.01 3.07 Wk4 3.04 1.96 2.23 2.98 Wk5 10.25 3.9 3.75 2.43 Wk6 5.54 2.88 3.29 2.7 Wk7 5.1 2.11 3.02 1.97 Wk8 3.87 3.04 1.98 2.45 2016 Wk9 3.47 4.1 2.56 5.7 Wk10 5 3.72 9.41 5.32 Wk11 16.75 8.62 2.19 3.14 Wk12 8.96 5.46 4.98 9.42 Wk13 12.85 5.47 3.46 6.81 Wk14 17.19 8.18 2.27 5.12 Wk15 24.35 11.3 11.21 7.07 Wk16 22.13 2.63 13.75 7.73 2017 Wk17 23.35 7.47 13.81 9.94 Wk18 10.43 4.37 8.21 5.59 Wk19 14.29 4.7 13 4.84 Wk20 21.55 4.05 31.87 8.77 Wk21 29.53 3.91 3.34 2.52 Wk22 30 5.84 8.98 5.83 Wk23 35.22 7.5 12.39 8.38 Wk24 30.15 5.7 12.44 7.37 2016 Wk25 2.46 4.65 16.53 5.45 Wk26 1.69 3.36 4.76 4.27 Wk27 2.91 1.72 1.27 1.28 Wk28 29 4 5.06 2.67 Wk29 24.63 5.91 31.76 7.47 Wk30 23.08 4.39 31.3 4.5 Wk31 6.59 7.27 7.73 8.98 Wk32 3.01 2.35 4.88 2.65 2017 Wk33 5.05 5.7 6.93 5.53 Wk34 10.11 3.45 9.91 3.73 Wk35 16.92 5.47 9.04 12.31 Wk36 17.54 5.35 10.11 5.75 Wk37 7.07 6.65 12.37 6.67 Wk38 5.91 4.05 8.74 5.71 Wk39 8.38 5.2 4.88 3.82 Wk40 7.11 6.11 6.34 4.07 III. RESULTS AND ANALYSIS A. Blocked Call Rate Analysis First, a correlation test is performed on the BCR results for both rainy and dry seasons to ascertain their nature of relationship existing between them. The correlation test is denoted with �� ranges from -1 to 1 [19]. �� = �� ��� � ��� ( �"�) (3) Engineering, Technology & Applied Science Research Vol. 9, No. 2, 2019, 3998-4001 4000 www.etasr.com Eli-Chukwu & Onoh: Experimental Study on the Impact of Weather Conditions on Wide Code … The result (rs=0.453) shows there is a slight positive relationship between the BCR for both rainy and dry seasons, such that a rise in rainy season BCR leads to a rise in dry season BCR: μ%�& '()* = 12.8955 , μ�%'()* = 8.7573 . A statistical approach for testing the hypothesis is used to ascertain if there is a difference in population means between the BCR during the rainy and dry seasons. Hypothesis: 3�: μ% 5 μ� = 0 3�: μ% 5 μ� 7 0 (4) Decision Rule: Accept H0 if 8��9 : 8( ;< �"=)�>(?) , otherwise reject. Test Statistic: 8��9 = (@;"@�)AB;��;(?) (5) 8��9 = 2.796 E 8( ;< �"=)�>(?) = 1.664 (6) Conclusion: There’s a statistical significant difference between the blocked call rate for the rainy and dry seasons. The graph in Figure 1 shows that the rate of blocked calls during rainy season is higher than the dry/Hammattan season. This result is obvious in weeks 21, 23 and 25 respectively. Fig. 1. Rainy and dry season BCRs B. Drop Call Rate Analysis Applying (3)-(5) the DCR we get rs=0.524. With rs there is a slight positive correlation between the DCRs of both rainy and dry seasons: μ%�& 'G)* = 4.798 and μ�%'G)* = 5.2085 , 8��9 = 51.129 E 8( ;< �"=)�>(?) =1.664. Conclusion: There’s no statistical significant difference between the DCRs of rainy and dry seasons. The graph in Figure 2 shows that there is a statistical equal rate of dropped calls during rainy and dry/Hammattan seasons. All through the test period, the average marginal difference (AMD) of the DCR was 51 K LM�G)* K 1. Fig. 2. Rainy and dry season DCRs C. Comparative KPI Analysis Friedman test is a non-parametric test of hypothesis approach for repeated measure analysis of variances which is used when the same parameter has been measured under different conditions on the same subjects [20]. Hypothesis: 3�: �*()* = �N()* = �*G)* = �NG)* = 0 3�: �*()* = �N()* 7 0 (6) Decision Rule: Accept H0 if OP�9Q� E O�&R and reject if otherwise [20]. S���9 = �=� *�� T ��� U(U<�) 5 3V(W X 1) ~ S�( ,U)�>(?) (7) OP�9Q� = 0.05 E O�&R. = 0 Conclusion: There’s no statistical significant difference between the BCR and DCR for rainy and dry seasons. D. Multiple Linear Regression Multiple linear regression is a statistical approach that models the relationship between two or more explanatory variables and a response variable by fitting a linear equation to the observed data [20]. All the values of the independent variable x are associated with the values of the dependent variable Y. The model is expressed as: Z& = �� X ��[&�X..X �\[&\ X ]& (8) Z& = 2023.437 5 0.084[� X 0.756[= X 0.415[^ X 0.555[_ where Z& =years, [� =rainy BCR, [= =dry BCR, [^ =rainy DCR and [_ =dry DCR. The above linear regression model can be used for prediction. IV. CONCLUSION In cellular network activities in Nigeria, before site integration, it is expected that each cell of the site meets the NCC standard of ≥98% for service accessibility and call completion rate which in turn depicts a ≤2% in the BCR and Engineering, Technology & Applied Science Research Vol. 9, No. 2, 2019, 3998-4001 4001 www.etasr.com Eli-Chukwu & Onoh: Experimental Study on the Impact of Weather Conditions on Wide Code … DCR. These results are achieved during seasonal changes. This paper points out the impact of varying weather conditions on WCDMA network performance. The results showed that during adverse weather conditions, the BCR and DCR rise greater than 8% and 4% respectively. Although with a slight relationship between the weather conditions, the weather conditions during the dry season have a better BCR of 8.76% than the rainy season with 12.89%. Calls tend to drop more during the dry season. A regression model was developed for predicting unknown network call statistics variables. Optimization actions that will protect against the effects of weather conditions on service accessibility and retainability should be considered in future studies. REFERENCES [1] T. S. Dalip, V. Kumar, “Effect of Environmental Parameters on GSM and GPS”, Indian Journal for Science and Technology. Vol. 7, No. 8, pp. 1183-1188, 2014 [2] S. H. Fang, Y. S Yang, “The impact of weather condition on Radio- Based Distance Estimation: A case study in GSM Network with Mobile Measurement”, IEEE Transactions on Vehicular Technology, Vol. 65, No. 8, pp. 6444-6453, 2015 [3] M. S. Yadnav, I. W. Sudiartha, “Simulation of broadcast level signal mobile station 3G Network rain condition”, Asia Pacific Conference on Multimedia and Broadcasting, Bali, Indonesia, November 17-19, 2016 [4] J. Luomal, I. Hakala, “Effects of Temperature and Humidity on Radio Signal Strength in Outdoor Wireless Sensor Networks”, Computer Science and Information Systems, Vol. 5, pp. 1247-1255, 2015 [5] M. Voznak, J. Rozhon, “Influence of Atmospheric Parameters on Speech Quality in GSM/UMTS”, International Journal of Mathematical Models and Methods in Applied Sciences, Vol. 6, pp. 575-582, 2012 [6] J. Rozhon, P. Blaha, M. Voznak, J, Skapa, “The Weather Impact on Speech Quality in GSM Network”, in: Computer Networks, Communications in Computer and Information Science, Vol. 291 Springer, 2014 [7] M. C. Kestwal, S. Joshi, L. S. Garia, “Prediction of Rain Attenuation and Impact of Rain in Wave Propagation at Microwave Frequency for Tropical Region”, International Journal of Microwave Science and Technology, Vol. 2014, Article ID 958498, 2014 [8] S. Sabu, S. Renjmol, D. Abhiram, B Premlet, “Effect of rainfall on cellular signal strength: A study on the variation of RSSI at user end of smartphone during rainfall 2017 IEEE Region 10 Symposium, Cochin, India, July 14-16, 2014 [9] B. Fong, P. B. Rapaiic, G. Y. Hong, “Effects of rain attenuation on wireless transmission of frame relay traffic”, The 8th International Conference on Communication Systems, Singapore, November 28, 2002 [10] J. Skapa, M. Dvorsky, L. Michalek, R. Sebesta, P. Blaha, “K-mean clustering and correlation analysis in recognition of weather impact on radio signal”, 35th International Conference on Telecommunications and Signal Processing, Prague, Czech Republic, July 3-4, 2012 [11] L. Ali, I. Alam, A. A. S. Syed, M. Yaqoob, “Various Meterological Parameters Effect on GSM Radio Signal Propagation for a Moderate Area”, 2017 International Conference on Frontiers of Information Technology, Islamabad, Pakistan, December 18-20, 2017 [12] P. Blaha, J. Rozhon, M. Voznak, J. Skapa, “Correlation between Speech Quality and Weather”, in: Soft Computing Models in Industrial and Environmental Applications. Advances in Intelligent Systems and Computing, Vol. 188, Springer, 2014 [13] L. A. R. da Silva Mello, E. Costa, R. S. L. de Souza, “Rain attenuation measurements at 15 and 18 GHz”, Electronics Letters, Vol. 38, No. 4, pp. 197-198, 2015 [14] C. Li, X. Luo, C. Zhang, X. Wang, “Sunny, Rainy, and Cloudy with a Chance of Mobile Promotion Effectiveness”, Marketing Science, Vol. 36, No. 5, pp. 762-779, 2017 [15] D. D. Dajah, N. Parfait, “A Consideration of Propagation Loss Models for GSM during Harmattan in N’djamena (Chad)”, International Journal of Computing and ICT Research, Vol. 4, No. 1, pp. 43-48, 2010 [16] N. H. Sabri, R. Umar, M. M. Shafie, S. N. A. S. Zafar, R. Mat, A. Sabri, Z. A. Ibrahim, “Correlation Analysis of Tropical Rainforest Climate Effect on Radio Signal Strength at KUSZA Observatory, Terengganu”, Advanced Science Letters, Vol. 23, No. 2, pp. 1268-1271, 2015 [17] A. F. Ismail, M. R. Islam, J. Din, A. R. Tharek, N. L. I. Jamaludin, “Investigation of rain fading on a 26GHz link in tropical climate”, 6th International Conference on Telecommunication Systems, Services, and Applications, Bali, Indonesia, October 20-12, 2011 [18] N. C. Eli-Chukwu, G. N. Onoh, “Improving Service Accessibility (CSSR) In GSM Network using an Intelligent Agent-Based Approach”, International Journal of Computer Engineering in Research Trends, Vol. 4, No. 11, pp. 478-486, 2017 [19] P. K. Sahu, S. R. Pal, A. K. Das, Estimation and Inferential Statistics, Springer, 2015 [20] M. R. Spiegel, L. J. Stephens, Theory and Problems of Statistics, McGraw-Hill, 1999