Energy-Efficient Power Management for O-RAN Base Stations Utilizing Pedestrian Flow Analytics and Non-Terrestrial Networks

Abstract

Recently, the open radio access network (O-RAN) architecture has been expected to enhance both the openness of network components and the intelligence of control functions as a promising RAN architecture for Beyond 5G (B5G)/6G networks. Meanwhile, the power consumption of base stations (BSs) in RAN is a serious problem that needs to be addressed owing to the recent increase in service types such as 4G-LTE, 5G, and local 5G, and it will be more remarkable in the future B5G era. However, a conventional RAN experiences energy wastage because it turns on the power of all BSs at all times, even in coverage areas that accommodate a small number of mobile terminals and low traffic. The O-RAN Alliance discusses the energy savings of BSs, but its standard specification lacks sufficient discussions on concrete models and protocols to realize highly energy-efficient power-on/off management of BSs. On the other hand, terrestrial network (TN) and non-terrestrial network (NTN) convergence has recently been considered in both academic research and standardization as an emerging technology for B5G networks. However, utilizing NTN capacities for BS power-on/off control of TN in the standard O-RAN architecture remains uninvestigated, although it has the potential to achieve higher energy efficiency. This study proposes a novel energyefficient power management architecture for O-RAN BSs. The proposed power management architecture extends the traditional standard O-RAN architecture such that the pedestrian flow analytics results and NTN capacities can be effectively utilized to obtain a higher energy-saving effect for O-RAN BSs. Consequently, the proposed power-on/off control reduces the power consumption of O-RAN BSs while maintaining the continuity of communications, bitrate, and other metrics. We performed numerical calculations using real datasets of pedestrian flows in regional mesh areas. As a result, we proved that the proposed architecture reduces power consumption by up to 40% when the NTN can accommodate UEs' traffic of approximately 400 Mbps. In addition, we implemented pedestrian flow analytics and power control functions in the controllers. We verified the feasibility of the functions by demonstrating the power-on/-off of an O-RAN BS using a mobile network testbed.