ZTE Corporation, a global leader in information and communication technology solutions, and China Mobile have jointly completed the research and laboratory verification of key technologies for their wireless network digital twin. This milestone includes multi-service twinning, wireless channel twinning and intelligent optimization decision-making based on ZTE's high-fidelity wireless network digital twin platform.
This successful verification lays the technological foundation for the joint cooperation of China Mobile's new generation AI open platform and ZTE's digital twin platform, leading the development of wireless network intelligence and promoting technology innovation to empower the digital transformation of operators.
The multi-service twinning technology utilizes temporal Generative Adversarial Networks (GAN) to replicate real video services such as iQiyi, Youku and Tencent within the digital twin environment. Notably, the packet mean error of the digital twin application remains within 10% when compared to real applications. This precision allows for evaluating the impact of real applications on user behaviors, network scheduling capabilities and network performance indicators with exceptional accuracy. This groundbreaking technology empowers operators to make informed decisions and effectively optimize their wireless networks.
The wireless channel twinning technology performs high-fidelity twinning of wireless channels in physical networks. By extracting specific channel characteristics, it further improves twinning precision. The RSRP error of the twinned wireless channel is controlled within 2 dB, and the errors of other major air interface indicators such as throughput are within 15%. Meanwhile, intelligent prediction technology is used for outdoor wireless channels, and the standard deviation between the predicted RSRP value and the measured RSRP value is within a range below 3 dB.
Intelligent optimization decisions are a critical approach to enhancing network intelligence. It offers researchers an approximation of a real wireless network twinning environment, where AI algorithms can be designed to make intelligent optimization decisions and improve network performance. Utilizing the digital twin platform, the output of intelligent optimization decisions results in an increase of more than 5% in downlink throughput and more than 25% in uplink throughput compared with manual optimization.
The verification results demonstrate that the precision of multi-service twinning and wireless channel twinning is comparable to the latest research results in the industry, providing a critical foundation for applying digital twin technology to real networks. As a result, digital twins significantly improve network intelligence, reduce repetitive costs and enhance network performance. Moreover, compared with manual optimization, digital twins offer technical support for subsequent intelligent network optimization.
Moving forward, China Mobile will deepen its collaboration with ZTE to explore network digital twin capabilities, including wireless network environment twinning, site twinning and terminal twinning. Driven by the demands of multi-scenario applications, this initiative aims to empower precise planning, efficient optimization and intelligent, simplified O&M activities by leveraging key digital twin technologies.
Through opening up capabilities for remote service calls, multi-scenario applications intend to build an open industry-academia-research integration ecosystem, co-create a new paradigm of intelligent services and initiate a novel digital twin cooperation model.