ZTE's Tangxue: 5G+AI for integrated communication and computing

The rapid advancement of Artificial Intelligence(AI) is revolutionizing various industries, ushering in a new era of innovation. Pioneering AI models such as ChatGPT and Sora by OpenAI, and Gemini by Google are at the forefront of this transformation, providing users with an expanded realm for creativity and imagination and opening up unprecedented possibilities. AI models are fundamentally reshaping social structures and daily routines by improving the ability to generate new ideas, solve complex problems, and automate routine tasks. Tang Xue, Vice President of ZTE, recently shared valuable insights and experiences with experts from diverse fields at the GSMA panel on Convergence of 5G & AI at the Edge during MWC2024 in Barcelona, Spain.

Tang Xue emphasized that the integration of 5G and AI is inevitable, ushering more possibilities for new business models and applications. As the edge node closest to users, performing AI processing at 5G network can effectively reduce data transmission delay, improve data processing efficiency, and reduce the cost of AI applications. While 5G networks provide large bandwidth, high speed, and low latency, making it ideal for supporting the connection and computing requirements for distributed AI computing. The integration of 5G and AI not only can achieve flexible collaboration of cloud and network, but also provide stronger intelligent capabilities, allowing 5G to adaptively optimize and adjust according to data characteristics and scenario requirements, improving network performance and efficiency.

“5G and AI integration, 5G-A new capability growth and new scenario exploration need a lot of computing power. Therefore, an integration of communication and computing infrastructure is required”, Tang said. ZTE 5G-A BBU is just this novel infrastructure, which is hardware ready for communication and computing integration, further explores computing power of RAN system, implements computing power orchestration adapting to full scenario requirements, including site computing power enhancement, inter-site computing power sharing and computing power integrated in computing force network, flexibly empowers B2C new services, B2B application in-depth development and low-altitude new economy. For V2X case, a built-in intelligent computing board can be added in 5G-A BBU for a heterogeneous computing power support and enabling more differentiated intelligent applications, such as 3^rd party platform deployment.

In addition, ZTE has been engaged in AI application innovation for a long time. RAN Composer is the industry's first native-AI based RAN intelligence solution. On top of 5G-A BBU, the integration of 5G and AI at edge nodes brings traditional network resource management revolution, achieves user-centered experience with precise resources allocation based on network serving capability, service requirements and UE capability, enabling supreme user experience, higher energy efficiency and O&M efficiency.

With 5G-A BBU, RAN Composer helps vertical industries achieve deterministic experience guarantees, adjusting resource allocation and scheduling strategies in real time through precise analysis of flows and packets. In WISCO (Wuhan Iron and Steel Corporation) case, remote control of crane is a typical application with large-bandwidth video to be transmitted in the uplink and highly reliable control instructions need to be transmitted in the downlink. A differentiated scheduling strategy will be performed based on AI learning and recognition of video data and control instructions. For example, video data identifies I frames (key frames of the picture) and uses smooth scheduling to avoid "I frame" collisions; control instructions use ML to predict the data packet sending for an accurate scheduling, which not only ensure the reliability of the control instructions, but also reduce the occupation of system resources.

5G-A BBU also supports the deployment of 3rd party applications, such as 5G-based V2X application. ZTE exhibited a 5G-based V2X case in cooperation with Tianyi Transportation this year. On an open road of 162 square kilometers, 5G connection and AI-based edge computing are used to achieve precise collaboration of roads, vehicles and clouds. The autonomous driving network reduces congestion rate by 20%, decreases carbon emissions by 12%, and cuts down accident rate by 8%.

In summary, the integration of 5G and AI presents significant opportunities and challenges for the development and monetization of 5G networks. It enables innovative applications and business scenarios, fully unleashing the value of 5G networks, such as deterministic connection guarantees, edge rendering, and V2X applications. These new applications and services can significantly increase the revenue of operators and service providers, serving B2C, B2B, and new economy development.

2024 is the first year of 5G-Advanced (5G-A) commercialization. Driven by new technologies, new services and new scenarios, the boundaries of communication networks continue to expand. 5G-A is not only the key of 5G development in the next 10 years, but also the key to shape future digital society. “Facing 5G-A, ZTE 5G-A BBU and RAN Composer are ready to unlock 5G value with high efficiency”, Tang added. For B2C scenario, connecting the virtual and real world for a digital life; For B2B scenario, facilitating 5G in core production and low-altitude economy for a digital society.