Prediction of Channel Stability to Optimize Signalling Overhead in 5G Networks using ML Algorithm

Abstract

As we are leaving in 5G era which promises us to give high throughput rate, better efficiency rate, and high bandwidth etc as its feature make headlines to give advanced and enhanced services in various sector whether it could be IOT, infrastructure, agriculture, and health care [1]. Giving well-grounded communication technology possesses a fundamental challenge for 5G system in both Core network and Radio Access Network (RAN) levels [2]. To get a better credibility at RAN level, base station (EnodeB or gNodeB in 4G and 5G respectively) should allocate proper amount of resource as per UE and select appropriate Modulation and scheme (MCS) to give better user experience. The amount of Physical Resource Blocks is directly proportional to channel condition at the User Equipment (UE) end [5]. The Base Station would generally know the real time channel strength of each UE, which then later provides accordingly resource to the UE, as UE would be sending periodically CQI to Base station. However, the periodic transmission of CQI would lead to signaling overhead which hampers the performance of RAN. Henceforth its foremost thing to average it or to optimize the CQI value to give a better user experience. Therefore, will be exploring four such machine learning algorithms (ANN, reinforcement learning, KNN and SVM) to optimize the CQI and would be analyzing which Algorithm will give better performance. Key Words: CQI, Machine Algorithm, RAN, Offloading, Handover, Base Station