An artificial intelligence enabled F-RAN testbed

F-RAN is regarded as a promising paradigm for mobile networks to alleviate the unprecedented traffic pressures and meet quality of service requirements of various 5G services with great flexibility. To make F-RAN work in a reliable, efficient, and smart way, AI-enabled F-RAN could be innovative in a number of directions: computing task offloading, resource management, dynamic beam selection, cross-layer design, energy saving and harvesting, mobility enhancement, and so on. In this article, an AI-enabled F-RAN testbed has been designed and implemented in a portable way based on OpenAirInterface, where an AI module is integrated into the F-RAN architecture. The AI module encapsulates the underlying operators of various machine learning frameworks to help a network make policies for different applications. Based on the proposed testbed, the F-RAN research community can easily analyze and evaluate their novel methods and quickly develop intelligent algorithms in a lab environment.