Towards 5G and beyond radio link diagnosis: Radio link failure prediction by using historical weather, link parameters

2022-04-01
Aktaş, Semih
Alemdar, Hande
Ergüt, Salih
Weather-related phenomena such as clouds, rain, snow affect the performance of radio links. To reduce the adverse effects of radio link failures’ on the user experience, mobile operators require intelligent monitoring systems to predict link failures and take actions before they happen. In this study, we show how machine learning can be used for prediction using a real-world telecom operator dataset. We propose a novel architecture to process time-series data and non-times-series data together in the same neural model to have better performance in predictions. We compare our model with the traditional approaches such as logistic regression (LR), support vector machines (SVM), and Long Short-term Memory (LSTM). Through experimental evaluations, we show that the F1-score of our proposed model is 0.638, whereas for the pure LSTM model it is 0.601. SVM and LR methods perform significantly worse with F1 scores of 0.455 and 0.105, respectively.
Computers & Electrical Engineering

Suggestions

ENGINEERING GEOLOGICAL CHARACTERIZATION OF THE ROCK MASSES AND THEIR EVALUATION BY SPATIAL ANALYSES, DETERMINATION OF THE ROCK SLOPE FAILURE SUSCEPTIBILITY ZONES AND HAZARD ASSESSMENT OF MUDURNU (BOLU)
Arslan Kelam, Arzu; Akgün, Haluk; Bobet, Antonio; Department of Geological Engineering (2022-7-20)
Mudurnu County, which is settled in a valley, is affected by regional rock slope instabilities especially due to precipitation, temperature changes, wind and earthquake activity. While units located on the eastern side of the valley are susceptible to rock falls, units at the western side are susceptible to planar, wedge or toppling failures. The presence of different rock mass properties makes Mudurnu an attractive and distinctive site for the investigation of rock mass failure. The rock slope failures ten...
Evaluation of the WRF & WRF-Hydro modeling system to better understand the hydrometeorological interactions over humid and semi-arid climate conditions
Düzenli, Eren; Yılmaz, Mustafa Tuğrul; Yücel, İsmail; Department of Civil Engineering (2022-10-7)
Changing climate causes extreme weather events, such as heavy precipitation and flash floods, to occur more destructive and frequently. Thus, it is worthwhile to perform a reliable prediction of hydrometeorological circumstances reasonably before extreme events occur. Similarly, robust estimation of future weather conditions would contribute significantly to long-term water management issues, one of the most crucial problems of our time. Accordingly, this study evaluates the skill of (un)coupled WRF/WRF-Hyd...
Cycle-Spinning GAN for Raindrop Removal from Images
Uzun, Ülkü; Temizel, Alptekin (2019-09-21)
Weather events such as rain, snow, and fog degrade the quality of images taken under these conditions. Enhancement of such images is critical for intelligent transport and outdoor surveillance systems. Generative Adversarial Networks (GAN) based methods have been shown to be promising for enhancing these images in recent years. In this study, we adapt the cycle-spinning technique to GAN for removal of raindrops. The experimental evaluation of the proposed method shows that the performance is improved in ter...
Evaluating the use of radar precipitation data in simulating a flood event in Terme Samsun Turkey
Akyürek, Sevda Zuhal (2016-10-10)
Floods caused by convective storms in mountainous regions are sensitive to the temporal and spatial variability of precipitation. In developing countries, researches and authorities working on this issue have to deal with lack of data. Even a dense raingauge network may not be able to state variability of the precipitation patterns especially in mountainous regions. Space-time estimates of rainfall from weather radar can be a remedy to represent pattern of the rainfall with some inaccuracy. In the field of ...
Evaluating the use of different precipitation datsets in flood modelling
Akyürek, Sevda Zuhal (2016-04-17)
Satellite based precipitation products, numerical weather prediction model precipitation forecasts and weather radar precipitation estimates can be a remedy for gauge sparse regions especially in flood forecasting studies. However, there is a strong need for evaluation of the performance and limitations of these estimates in hydrology. This study compares the Hydro-Estimator precipitation product, Weather Research and Forecasting (WRF) model precipitation and weather radar values with gauge data in Samsun-T...
Citation Formats
S. Aktaş, H. Alemdar, and S. Ergüt, “Towards 5G and beyond radio link diagnosis: Radio link failure prediction by using historical weather, link parameters,” Computers & Electrical Engineering, vol. 99, pp. 1–11, 2022, Accessed: 00, 2022. [Online]. Available: https://hdl.handle.net/11511/96033.