A Novel Front-haul Bandwidth Compression Method for RAN Systems

Abstract

Recently, there has been a significant increase in users as well as user data requirements in mobile communications. This is attributed to advances in mobile communication systems and networking, along with the advent of fifth generation (5G) mobile systems. As a result, front haul data compression techniques have become necessary to meet QoS requirements. In this paper, we resort to contemporary machine learning techniques and provide algorithms to, respectively, dynamically predict and compress the front haul data. The proposed scheme involves evaluating the Error Vector Magnitude (EVM) metric and comparing the performance with existing schemes. Furthermore, these algorithms can be deployed on contemporary C-RAN as well as O-RAN architectures. From simulations, we are able to demonstrate a compression of about 65%.