Home About Login Current Archives Announcements Editorial Board
Submit Now For Authors Call for Submissions Statistics Contact
Home > Archives > Volume 20, No 11 (2022) > Article

DOI: 10.14704/nq.2022.20.11.NQ66265

Proactive Energy Saving Technique for Cellular Base Station

Mamta Punjabi, Gend Lal Prajapati

Abstract

The increasing demand of communication infrastructure is increasing the power consumption in a cellular base station. Smart management of running devices in base stations can reduce the demand of power supply and making a step towards Green Information Technology (IT). In this paper we provide a study on green IT which helps to optimize energy consumption in cellular network base stations. In this context we organize this paper in three parts: first provide a review of recent prediction techniques, dataset used, and employed applications. Next we introduced a prediction technique for predicting future workload on a base station. This predictive model involves a modified k-means clustering for behaviour based traffic clustering. Then the clustered cellular traffic learned with the three Machine Learning (ML) techniques namely a k-Nearest Neighbour (k-NN) lazy learner, regression based learning technique, and a deep learning technique Long Short Term Memory (LSTM) to learn the traffic pattern. Using predicted workload, the algorithm is tried to reduce the energy requirements of the base stations. In order to validate the proposed model, we utilize the 4G LTE cellular network traffic dataset. Additionally to implement the model the Google Colab service has been used. According to experimental analysis we have found Deep learning based architecture provide very poor accuracy for this model and only able to produce up to 50% of accuracy. On the other hand regression based learning technique provides second winning algorithm. Additionally, the lazy learning technique is able to provide the acceptable higher accuracy. The implementation of the energy preserving algorithm with k-NN confirms the model can preserve the energy maximum 20-28% of the cell towers.

Keywords

Machine Learning, Lazy Learning, Cellular Network Traffic Prediction, Time Series Problem, Enhancing Performance Of Prediction.

Full Text

PDF

References

?>