Small Cell Management in Cellular Networks based on User Density Prediction

Inhalt:
Mobile communication is in its fifth generation (5G) and the demand for mobile broadband services is ever increasing. It is a major challenge for the network operators to manage the surge in data traffic volume. When there is an upswing in the number of users to be served by a cell then such a basestation will become overloaded. There are certain sites of interest where a spurt in mobile user density is more likely. Venues such as, library, shopping malls, public halls, concerts, stadium, etc., are few examples. In order to relieve such high load, small cells (SC) can be deployed and appropriately activated within the coverage area of macro cells (MC) based on user crowd. A pro-active small cell management framework is proposed in this paper based on user density prediction. Random forest regression is employed to predict the user densities throughout the day at the point of interest based on user history (features such as semester information, time of the day, temperature, etc.,). This is used in tandem to activate or deactivate the small cell, instead of keeping the small cell always on. Simulations performed demonstrate significant improvement in the number of users served while using the SC management framework.