O-RAN Intelligent Application for Cellular Mobility Management

Abstract

Machine Learning (ML) is gaining a more important role in cellular networks. Mobility management can be improved using ML due to its complexity and criticality to network performance. Managing mobility while maintaining the network connection is very important because the user's movement may disrupt their data transmission. In some cases, the traditional mobility management algorithm is not reliable enough. Here is where ML may provide a more intelligent mobility management algorithm. ML implementation in the cellular network requires a major modification in the network element software logic and architecture. This may cause difficulties in real-world implementation. Open Radio Access Network (O-RAN) consortium provides a modular solution to implement ML algorithms by adding an optional Radio Intelligent Controller (RIC) to host the intelligent application without major modification to the existing network elements. Using this RIC, a lot of intelligent use cases can be implemented modularly in the network. In this paper, we prove that the ML algorithm can be used to improve mobility management in some particular cases. As the experiment results, we demonstrated that machine learning has superior performance compared to the traditional handover algorithm.