Repository logo

Doctoral Degrees (Electronic Engineering)

Permanent URI for this collectionhttps://hdl.handle.net/10413/6867

Browse

Browsing Doctoral Degrees (Electronic Engineering) by Subject "Artificial neural networks."

Filter results by typing the first few letters
Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    The impact of visibility range and atmospheric turbulence on free space optical link performance in South Africa.
    (2022) Layioye, Okikiade Adewale.; Afullo, Thomas Joachim Odhiambo.; Owolawi, Pius Adewale.
    In the recent years, the development of 5G and Massive Internet of Things (MIoT) technologies are fast increasing regularly. The high demand for a back-up and complimentary link to the existing conventional transmission systems (such as RF technology) especially for the “last-mile” phenomenon has increased significantly. Therefore, this has brought about a persistent requirement for a better and free spectrum availability with a higher data transfer rate and larger bandwidth, such as Free Space Optics (FSO) technology using very high frequency (194 𝑇𝐻𝑧−545 𝑇𝐻𝑧) transmission system. There is currently unavailable comprehensive information that would enable the design of FSO networks for various regions of South Africa based on the impact of certain weather parameters such as visibility range (mainly in terms of fog and haze) and atmospheric turbulence (in terms of Refractive Index Structure Parameter (RISP)) on FSO link performance. The components of the first part of this work include Visibility Range Distribution (VRD) modeling using suitable probability density function (PDF) models, and prediction of the expected optical attenuation due to scattering and its cumulative distribution and modeling. The VRD modelling performed in this work, proposed various location-based PDF models, and it was suggested that the Generalized Pareto distribution model best suited the distributions of visibility in all the cities. The result of this work showed that the optical attenuation due to scattering within the coastal and near-coastal areas could reach as high as 169 𝑑𝐵/𝑘𝑚 or more, while in the non-coastal areas it varies between 34 𝑑𝐵/𝑘𝑚 and 169 𝑑𝐵/𝑘𝑚, which suggests significant atmospheric effects on the FSO link, mostly during the winter period. The BER performance analysis was performed and suitable mitigating techniques (such as 4 × 4 MIMO with BPSK and L-PPM schemes) were suggested in this work. The general two-term exponential distribution model provided a good fit to the cumulative distribution of the atmospheric attenuation due to scattering for all the locations. In order to ascertain how atmospheric variables contribute or affect the visibility range, which in turn determines the level of attenuation due to scattering, a time series prediction of visibility using Artificial Neural Network (ANN) technique was investigated, where an average reliability of about 83 % was achieved for all the stations considered. This suggests that climatic parameters highly correlate to visibility when they are all combined together, and this gave significant predictions which will enable FSO officials to develop and maintain a strategic plan for the future years. The modules of the second part of this work encompass the determination of the Atmospheric Turbulence Level (ATL) for each of the locations in terms of RISP (𝐶𝑛2) and its equivalent scintillation index, and then the estimation of the optical attenuation due to scintillation. The cumulative distributions of the optical attenuation due to scintillation and its modeling were also carried out. This research work has been able to achieve the prediction of the ground turbulence strength (through the US-Army Research Laboratory (US-ARL) Model) in terms of RISP using climatic data. In an attempt to provide a more reliable study into the atmospheric turbulence strength within South Africa, this work explores the characteristic behavior of several meteorological variables and other thermodynamic properties such as inner and outer characteristic scales, Monin-Obhukov length, potential temperature gradient, bulk wind shear and so on. According to the predicted RISP from meteorological variables (such as temperature, relative humidity, pressure, wind speed, water vapour, and altitude), location-based and general attenuation due to scintillation models were developed for South Africa to estimate the optical attenuation. The attenuation due to scintillation results show that the summer and autumn seasons have higher ATL, where January, February and December have the highest mean RISP across all the locations under study. Also, the comparison of the monthly averages of the estimated attenuations revealed that at 850 nm more atmospheric turbulence with specific attenuations between 21.04 𝑑𝐵/𝑘𝑚 and 24.45 𝑑𝐵/𝑘𝑚 were observed in the coastal and near-coastal areas than in the non-coastal areas. The study proposes the two-term Sum of Sine distribution model for the cumulative distribution of the optical attenuation based on scintillation, which should be adopted for South Africa. The obtained results in this work for the contributions of scattering and turbulence to the optical link, and the design of the link budget will serve as the major criteria parameters to further compare the outcomes of these results with that of the available terrestrial FSO systems and other conventional transmission systems like RF systems.
  • Thumbnail Image
    Item
    Modelling of propagation path loss using adaptive hybrid artificial neural network approach for outdoor environments.
    (2018) Ebhota, Virginia Chika.; Srivastava, Viranjay Mohan.; Isabona, Joseph.
    Prediction of signal power loss between transmitter and receiver with minimal error is an important issue in telecommunication network planning and optimization process. Some of the basic available conventional models in literature for signal power loss prediction includes the Free space, Lee, COST 234 Hata, Hata, Walficsh- Bertoni, Walficsh-Ikegami, dominant path and ITU models. But, due to poor prediction accuracy and lack of computational efficiency of these traditional models with propagated signal data in different cellular network environments, many researchers have shifted their focus to the domain of Artificial Neural Networks (ANNs) models. Different neural network architectures and models exist in literature, but the most popular one among them is the Multi-Layer Perceptron (MLP) ANN which can be attributed to its superb architecture and comparably clear algorithm. Though standard MLP networks have been employed to model and predict different signal data, they suffer due to the following fundamental drawbacks. Firstly, conventional MLP networks perform poorly in handling noisy data. Also, MLP networks lack capabilities in dealing with incoherence datasets which contracts with smoothness. Firstly, in this work, an adaptive neural network predictor which combines MLP and Adaptive Linear Element (ADALINE) is developed for enhanced signal power prediction. This is followed with a resourceful predictive model, built on MLP network with vector order statistic filter based pre-processing technique for improved prediction of measured signal power loss in different micro-cellular urban environments. The prediction accuracy of the proposed hybrid adaptive neural network predictor has been tested and evaluated using experimental field strength data acquired from Long Term Evolution (LTE) radio network environment with mixed residential, commercial and cluttered building structures. By means of first order statistical performance evaluation metrics using Correlation Coefficient, Root Mean Squared Error, Standard Deviation and Mean Absolute Error, the proposed adaptive hybrid approach provides a better prediction accuracy compared to the conventional MLP ANN prediction approach. The superior performance of the hybrid neural predictor can be attributed to its capability to learn, adaptively respond and predict the fluctuating patterns of the reference propagation loss data during training.