AI-Powered Platform: Precise Network Energy-Saving Management
The energy saving platform, based on ZTE VMAX system, automatically obtains network data including engineering parameters, MR measurement data, historical performance KPI data, and neighbor cell relationships from the RAN domain through scenario-based open APIs.
—Based on big data analysis, the platform automatically identifies network energy saving scenarios, predicts network traffic trends, such as traffic busy/idle hours and areas, traffic/energy consumption trends, identifies co-coverage cells and automatically generates energy saving policies. 
—The system automatically delivers energy saving policies and implements network-level intelligent energy saving policy management and site energy-saving coordinated scheduling. 
—Real-time monitoring of impact on network KPIs and energy saving benefits is achieved, allowing visualized management of energy saving benefits of mobile networks.
Large-scale deployments of it have shown an increase in power-saving activation rate by more than 80%. 
AI-driven network energy saving solution is not a particularly new concept. Equipment vendors have started to offer AI-driven energy-saving solutions as an extension to existing network management platforms to enable more precise energy savings based on traffic and other site related conditions, improving efficiency and reducing the manpower required. 

AI-Powered Site: Multidimensional Service-Aware Energy Saving
AI-driven service awareness has been taken into consideration since 2020, which exploits the differences in the energy efficiency of different types of services to deliver certain services to the most energy-efficient network, helping achieve the most efficient energy usage.
However, the multidimensional complexity of 5G is posing tremendous challenges to the transition from network-centered to user-centered. Different network resources, such as bands, modes, carriers, frame structures, and beams can be combined flexibly; and it's very complex to predict and select the resource combination capability in the violently changing radio environment. Meanwhile, due to industrial evolution and diverse application scenarios, terminal capabilities vary greatly in terms of support for system, mode, feature and 2B/2C services, which is an important constraint for scenario-based and personalized services. Thus, only service type based user steering is not enough. 
With the improvement of computing power of base station, more intelligent strategies can be introduced to base station and equipment. Cross-domain information that may affect energy consumption can all be obtained automatically by base station to make the user steering more precise.
Network service capability knowledge library introduced by ZTE Radio Composer solution will be applied and combined with service pilot to balance between network energy saving and user experience. Built on sub-cells of various scenario-based capabilities, the library predicts different cells' network service capabilities at different frequency layers through learning the history data, and combining with real-time information (e.g. service type, rate request and user type). 

5G Terminal Energy Savings
A report from GSMA Intelligence highlights that one mobile connection required an average of 14.8 kWh of energy during the 12 months of operation. As of the end of September 2021, the number of 5G terminals had reached 445 million in China, which means over 6.5 billion kWh will be consumed yearly. At the same time, the subscribers are constantly complaining about the heating problem and battery life of 5G terminals.
For better 5G experience and lower personal carbon footprint, PowerPilot supports more technologies to assist mobile phones working in an energy-saving state for a longer period of time, and the battery life can be extended by 20%. 
Bandwidth part (BWP) combined with uplink smart pre-scheduling and inactive state of radio resource control (RRC Inactive) reduces the amount of signaling required during state transitions, making it possible to significantly lower both latency and battery consumption. The access latency was shortened by up to three times in a trial with China Unicom. This shortened time lag will have a big impact on user experience in applications such as cloud gaming where multi-player interactions require 20–30 ms end-to-end latency, enable the battery savings of up to 30% and improve the battery life by over 20%.
PowerPilot was awarded "Outstanding Catalyst—Impact for Society" at TMForum 2020 and "Innovative Breakthrough in Mobile Technology Award" at GTI 2021. Till now, the solution has been deployed in 30 networks over 800,000 sites. In typical network configurations, the power consumption of 5G base stations can be reduced by 20%–25%, and the emission of about two million kg carbon dioxide can be avoided for every 1000 base stations in one year. The average shutdown duration is 9.88 hours, which is a 60% improvement compared with that when the feature is manually enabled. 
We all have a role to play if we are going to meet the challenges of the global climate crisis, and we do believe we can make some contribution to the mobile net zero target.