Search

Dust particles and sand storms can cause attenuation and cross-polarization of electromagnetic wave propagation, especially at high frequencies above 10 GHz. Dust attenuation has been the focus of many research works, mainly with the deployment of a 5G wireless network in the FR-2 band (mmWave band, 23–53 GHz with TDD). This has led to the development of novel models to accurately predict and estimate attenuation. However, the existing review works have not adequately provided extensive taxonomies for these models to show the state-of-art and future research directions. This paper aims to bridge this gap by providing a comprehensive review of all electromagnetic scattering models in terms of their strengths, weaknesses, and applications. Lessons learned from the detailed survey have been stated and discussed extensively. Key findings from this review indicate that all the models developed were limited to the region where they were developed, with frequency and visibility levels as the two main parameters. The survey across regions showed no model was developed for Region 2, including the Americas, Greenland, and some of the eastern Pacific Islands. Among the dry regions of the globe, where dust and sand storms can occur either occasionally or frequently, it can be seen that only a few parts of these desert regions of Africa (Region 1) and Asia (Region 3) have been considered by authors for the development of prediction models for attenuation due to dust storms. Thus, this also shows the limitations of the overall deterministic models and presents the crucial need to develop new models or modify existing models to accurately predict dust attenuation in other regions, particularly in Africa.

Features such as the communication scheme, energy awareness, and task distribution amongst others are the key component that characterizes the Flying Ad-hoc Network (FANET). The operational efficiency in FANET surveying a specific region is affected by the nature of the UAVs' node placement, routing protocol, energy-aware task distribution, and node interaction amongst others. In this paper, Drone 1 (D1), Master Drone (DM), and Drone 2 (D2) were used to survey a pipeline of length 12.2 m. This paper aims at minimising energy use by drones during surveillance using energy-aware node exchange technique, task interaction and distribution scheme for each UAV. Due to fast energy depletion of DM due to packets aggregation, its election is based on the UAV with the highest energy before take-off. For two different simulations, 14,697.0 J and 14,836.6 J were obtained for DM. To avoid system failure due to fast energy loss of DM, the drones swapped positions and status. First swapping command comes up when DM loses 50% of its energy, while the second command occurs when it further loses 15%. Return to base threshold energy is computed for the three UAVs to avoid crash due to insufficient energy during surveillance. DM returns to base threshold energy for both single and double swapping simulation were 658.105 J and 652.456 J respectively. From the results obtained the algorithms were able to exchange nodes to maximize energy usage and perform an interaction-based task distribution for cooperative task sharing during surveillance. This translates into longer surveillance time and effective telemetry data aggregation.

Several orthodox approaches, such as empirical methods and deterministic methods, had earlier been used for the prediction of path loss in wireless communication systems. These approaches are either inefficient or complex. Robustness and performance motivated the adoption of machine learning for modeling path loss in wireless communication systems in place of traditional modeling schemes. Surveys on modeling path loss in communication systems exist in the literature; however, emerging deep learning architectures in-depth analysis, machine learning taxonomies related to path loss, and in-depth analysis on feature engineering in modeling path loss are missing in the already published surveys. To fill this existing gap, a survey of machine learning modeling for path loss in wireless communication systems is conducted to resolve the outlined issues, hence making this survey unique. Synthesis and analysis of deep learning architectures to solve path loss problems in communication systems are hereby presented. New taxonomy – deep learning, nature-inspired meta-heuristic algorithms, and shallow algorithms approach to path loss modeling have been created. Analysis of feature engineering in path loss modeling is exploited. Lastly, challenges militating against the full potential of modeling path loss based on machine learning are highlighted and discussed. Alternative approaches for resolving the outlined challenges are also presented in the survey to help in designing more practical applications in the future.

Path loss propagation models are critically needed for optimum planning and deployment of wireless communication networks. However, the complexity exhibited by the propagated signals makes the prediction of the received losses difficult in built-up environments. There is however a new paradigm shift towards the application of computational methodologies, such as the Artificial Neural Networks (ANN), for multi-band path loss prediction. In this paper, we have developed a new ANN-based model for path loss prediction. The model was developed using large scale path loss data collected across 485 base stations in 6 urban cities of Nigeria, West Africa. The data collection, which spanned a period of 9 years, were taken over open areas, sub urban and urban environments, and the bands considered were 89.3 MHz, 103.5 MHz, 203.25MHz, 429.25 MHz, 529.25MHz, 615.25MHz, 629.25MHz, 900MHz, 1800 MHz and 2100MHz. The developed model was validated using independent path loss data across different frequencies, environments; and, distances and the results were compared with the popular empirical models such as Hata, COST 231 and Egli models, and to previously published ANN-based multi-frequency models. The global performance of 4.81 dB was obtained in terms of RMSE value with an R-value of 0.96, thus outperforming the existing ANN-based path loss models that were developed for multiple frequencies. Based on these findings, the proposed model can be deployed across all categories since the average RMSE values are all within the acceptable thresholds. Furthermore, the model is multi-frequency, thus will be suitable for multiple and complex environments, and usable for both short- and long-range wireless communication networks