Dual-Intelligent Network—A New Direction for Network Operation and Maintenance

Overview of Dual-Intelligence Network
With the large-scale commercial use of 5G networks, improving the utilization of 5G network resources while accurately ensuring differentiated user experience has gradually become the core proposition of network optimization, evolution and development. Operators are eager to improve 5G utilization, and often adopt a one-size-fits-all approach to adjust the interoperability of 5G->4G across the entire network to a lower threshold, thereby expanding the coverage of 5G cells. However, different sites have different wireless environments, and a unified interoperability strategy is difficult to adapt to all base stations. Therefore, the innovative solution of Dual-Intelligence Network came into being.

Dual-intelligent network, that is, network-level intelligent interconnection parameter self-optimization + user-level intelligent user orchestration. Intelligent interconnection parameter adjustment ensures network performance and cell-level perception to improve 5G retention ratio/diversion ratio; at the same time, intelligent user orchestration migrates users with poor perception to appropriate heterogeneous frequency or heterogeneous system cells to ensure user experience and perception.

Dual-intelligent network key technologies
Dual-intelligent network solutions mainly have two key technologies


Network-level intelligent interconnection parameter self-optimization technology: Based on the evaluation of various basic indicators of 5G cells and user perception indicators, the interconnection parameters from 5G to 4G are automatically adjusted. Under the premise of stable network performance indicators, the coverage of 5G cells is expanded as much as possible, the 5G retention rate is improved, and the upper limit of network service capacity is increased.



User-level intelligent user orchestration technology: Intelligent user orchestration is enabled at the same time to monitor user experience in near real time, and users with poor experience are migrated to other 5G frequencies or 4G cells, providing a bottom-line guarantee for intelligent intercommunication parameter adjustment. At the same time, it can also react to network parameter optimization and promote network performance improvement.
Intelligent intercommunication parameter self-optimization technology


Intelligent intercommunication parameter self-optimization technology, based on the evaluation of key basic indicators and user perception indicators of 5G cells, automatically adjusts the intercommunication parameter thresholds (connected state A2 event threshold and TTT, idle state non-co-frequency cell reselection measurement threshold), to achieve a one-site one-policy interoperability strategy, avoiding the expansion of the coverage of 5G cells and maximizing 55G.



As shown in Figure 3, during the "adjustment evaluation time window" evaluation time, the perception indicators meet the threshold requirements at the same time. By starting the heterogeneous system measurement threshold A2 in the service cell to automatically adjust downward, or adjusting the TTT of the A2 event, it is more difficult for users to trigger heterogeneous system intercommunication to 4G. After the parameter adjustment, observe the changes in the cell indicators. During the "fast post-evaluation" evaluation time and the "post-evaluation time window" evaluation time, if the absolute value of the perception indicator degradation is less than the set threshold, the automatic adjustment will take effect, otherwise the parameters will be rolled back to before the adjustment.


The lower limit of the heterogeneous system A2 should be greater than that of the blind redirection A2, and greater than or equal to the "minimum threshold of heterogeneous system A2 for residence ratio improvement". If you also choose to adjust the TTT of the A2 event, the TTT of the A2 event will be increased after the heterogeneous system A2 drops to the minimum value.

Note: If the cell intercommunication parameters are manually modified during the evaluation cycle, the automatic evaluation will be restarted.

Intelligent user orchestration technology
Intelligent user orchestration can monitor the service experience of each user in real time. When the experience deteriorates, it comprehensively considers the terminal capabilities, business requirements, and service capabilities that the network can provide (network heterogeneous frequency/system coverage, cell load conditions at each frequency layer, channel quality conditions, etc.), and combines the AI ​​wireless fingerprint grid library to quickly migrate users to the optimal target frequency layer and cell.


The current user orchestration technology includes the following functions:

Source cell user experience evaluation function: Uplink enhanced quality cut (SingleRBSINR+blocking rate joint judgment). Downlink enhanced quality cut ("downlink quality cut+downlink perceived low rate+downlink blocking" joint judgment).

Target cell precise selection: Neighboring cell load judgment (RRC user ratio, uplink/downlink PRB utilization). Neighboring cell downlink coverage and quality judgment (downlink RSRP, RSRQ, SINR). Neighboring cell uplink quality judgment (SingleRBSINR). Neighboring cell downlink air interface capacity judgment (average air interface capacity rate estimation).



Fast migration function: Based on fingerprint grid, the function of measuring heterogeneous frequency data quality is realized. Based on fingerprint grid, the function of measuring heterogeneous system data quality is realized.
Practical application of the solution
The dual-intelligent network solution has achieved a deployment scale of more than 10,000 sites in the three major domestic operators.

Taking a key city as an example, the network configuration of 5G->4G threshold and TTT before deployment is (-112dbm, 320ms), and the intercommunication self-optimization is activated and the intelligent user orchestration function is deployed simultaneously to ensure user-level business performance perception.


After the deployment of the dual-intelligent network, the operation results are as follows:

There are 1,148 NR cells in the promotion area, 31.27% of which maintain the original network configuration, and 68.73% of which successfully reach the lower limit. The specific operation status is shown in the table below.

The effect of the dual-intelligent network: While improving the 5G diversion ratio, it can ensure the stability of the overall user perception and performance indicators.



Compared with the same period in April, the 5G traffic in the dual-intelligence network promotion area from May 4 to 8 increased by 6.70%, and the 5G diversion ratio increased by 1.75%; the 5G traffic in the control area during the same period increased by 5.04%, and the 5G diversion ratio increased by 1.24%. Therefore, the actual net growth rate of 5G traffic in this dual-intelligence promotion is 1.66%, and the net gain of 5G diversion ratio is 0.51% (excluding natural growth).


The success rate of 5G inter-system switching in the promotion area improved by 0.15%, and other performance indicators remained stable, with no new complaints during the pilot period. Among the perception indicators, the CQI yield rate improved by 0.23%, the RRC reconstruction ratio improved by 0.06%, the uplink high-path loss ratio improved by 0.53%, the uplink UEUE rate to low rate improved by 0.017%, the uplink high-path loss ratio improved by 0.53%, and the uplink UEUE rate to low rate improved by 9%.



Wireless networks have evolved from 1G to 5G, and maintenance methods have continued to improve. Manual network optimization is inefficient, and quality and timeliness cannot be guaranteed. Network intelligent automation is the future development direction of network optimization. The network dynamically adjusts to its optimal state through continuous self-evaluation and self-optimization, and selects the best network for users through near-real-time user experience perception.