Operators have long been looking for ways to improve user experience and enhance utilization of expensive yet limited spectrum, whereas with the introduction of 5G, the coexistence and co-evolution of multi-band and multi-mode networking has greatly increased network complexity. As a key infrastructure of the digital society, 5G network not only serves individual users, but also meets the digital transformation needs of industries. The increasingly diversified vertical application and consumer services have further deepened the contradiction between “one-fit-all” network policy and differentiated user experience requirements. How to develop new services and deliver user-centric experiences for industry applications and consumers within limited network resources is of great importance to operators.
Based on the understanding of network development and in-depth research, ZTE innovatively proposed Radio Composer, the industry's first AI-driven context-aware customized experience solution. Starting from multi-layer of whole network, Radio Composer enables multi-dimensional perception and machine learning of UE capability, service characteristic, network characteristic and inter-domain information, manages and optimizes radio resources to meet service requirements in a dynamic way by changing the traditional "best effort" mechanisms to “always optimal", implements best mapping between service requirement and radio resources. It empowers user-centric policy to achieve a near real-time, customized user experience, facilitates best user experience and highest network efficiency, expands 5G users, unlocks network potential and promotes 5G industry and ecosystem development.
ZTE Radio Composer takes the lead to use BBU native Intelligence as the key driver of wireless network intelligence, together with advanced radio resource allocation algorithm to empower consumers and industry applications with first-class-like performance without the necessity of extra hardware, enables network policy shifting from“one-size-fit-all”to“context-aware“. The AI training is implemented when CPU working at low load in order to minimize impact on legacy user experience, and computing capability coordination will be activated among neighbor cells as the development of 5G users. The BBU native intelligence enables edge computing, which makes service application as close to access as possible. This facilitates the implementation of low-latency services, like gaming, AR/VR, which exploits more application scenarios and opportunities, accelerates the monetization of 5G network and improves operator’s revenue.
Radio Composer features flexible radio resource scheduling via network orchestration and user orchestration, among which user orchestration raises the upper limit of user experience while network orchestration further extends the upper limit of network serving capability.
Network orchestration
Network orchestration targets at a future-oriented network design, enables spectrum, frame structure and beam pattern adaptive adjustment according to intelligent traffic prediction and interference avoidance on a cluster basis, implements accurate mapping between variable network serving capabilities and prediction of traffic pattern to improve network payload and reduce O&M cost. According to field test result, it increases network payload by 32% on average and reduces O&M cost by 46%.
For spectrum with DSS features such as 2.1GHz, though DSS is a cost-effective way to have 5G coverage based on legacy FDD equipment, DSS NR performance is severely limited due to the interference between LTE and NR. In particular, DSS NR performance degrades due to the intra-cell LTE CRS and neighbor-cell LTE CRS interference, which has been the common pain-point among global operators. Radio Composer opens a door for DSS to improve performance and deploy in large scale. With intelligent traffic prediction on a cluster basis, the intelligent traffic steering of 2.1GHz LTE traffic to the co-coverage 1.8GHz cell can be implemented without affecting LTE user experience, thereby making LTE shutdown in the 2.1GHz DSS cell possible when LTE traffic been fully carried by the co-coverage 1.8GHz cell. Then there is only NR in the DSS cell without LTE CRS interference, an intelligent interference avoidance will be enabled to solve the interference between DSS cell and non-DSS cell, greatly improving DSS NR performance.
With the development of uplink services in both 2C and 2B scenarios, such as live webcast and high-definition video, large uplink bandwidth is required whereas current 5G network default frame structure is high speed downlink focused, like 3D1U, which limits the 5G network serving capability on the uplink. Radio Composer enables flexible frame structure adaption according to uplink and downlink service requirements, by adjusting the frame structure from 3D1U to 1D3U when there is large uplink demand, and from 1D3U to 3D1U when downlink is the major requirement. Comparing with manual frame structure adjustment, it not only improves network efficiency, but also saves operation and maintenance costs.
User Orchestration
User orchestration targets at smooth user experience with user steering and user scheduling, which achieve flexible steering and scheduling among multiple frequencies in the network based on accurate predictions of network service capabilities and precise recognition of service requirement, combined with terminal capabilities. In field network, it improves cell-edge user average throughput by more than 300%, reduces handover delay by more than 48% and decreases low-speed NR user ratio by 51%.
For video streaming services, users can enjoy 1080p video in good coverage area while degrading to 360p in poor. With Radio Composer, user will be steered to the cell with best experience when the experience is going to slide, therefore, smooth user experience can be ensured based on source cell experience evaluation and target cell experience prediction according to machine learning. In addition, the steering can be implemented in the shortest latency without waiting for measurement feedback. In a summary, Radio Composer improves user experience and saves video buffer delay and first frame playback delay.
For slicing service, some vertical applications have critical requirements on delay or reliability, such as smart manufacturing. To manage this in a traditional way, more resources would be allocated to guarantee the SLA requirement in a fixed manner, resulting in resource waste. With Radio Composer, the slicing resource can be managed in a dynamic way via machine learning of traffic pattern in daily or weekly basis. Consequently, with Radio Composer, 2 or more slices can be configured comparing with 1 fixed slice within the same network resources, so as to improve the spectrum utilization efficiency.
For vertical application terminals with different location characteristics, such as drone for inspection service and differential protection service, the traditional policy is to use a unified priority to reside on the high-frequency network with 5G capability. With Radio Composer, the drone inspection service can be accurately guided to the low band for wide coverage while the experience is also on-demand, and differential protection with fixed-location terminal are preferentially guided to high-frequency networks to better meet the differentiated services requirements, consequently improving diversified user demands.
Future-oriented evolution
Radio Composer will continue to evolve and grow in response to different stages of network development and differentiated service demands to exploit the best way to combine new service and application scenario with best experience and optimal efficiency. In addition, Radio Composer can be driven by more intentions, such as energy-saving intentions for carbon neutrality and the comprehensive integration of multiple intentions, like experience, efficiency and energy efficiency. With the introduction of cross-domain information, the joint orchestration of consumers and industry segments will be further strengthened to implement much flexible coordination of user experience, network efficiency and other intentions within a single network, so as to facilitate society digital transformation with 5G and promote prosperity of 5G business.