PARTNER FEATURE: Wireless network communication enters the high-speed growth phase. Operators pay more and more attention to network development quality and market expansion. With the increase of the number of online users and the continuous development of network services, the network load is facing challenges that affect user experience and brand effects. The customer faces the pressure of ROI decline and poor user experience, and the regular planning and expansion cannot meet customer expectations. Based on customer requirements, comprehensive network insight and prediction makes full use of network resources and precise network planning, ensuring user experience becomes the trend of operator customer demands.

In 2G, 3G and early 4G wireless network construction, limited by technical conditions such as tools, algorithms, etc., wireless network planning is mainly based on simulation and prediction. In 5G network, because of the introduction of mass MIMO and other key technologies, the traditional network planning method has been unable to meet the needs of new technologies for network planning. The technical innovation of wireless network planning is a perquisite. Therefore, the whole communication industry is constantly exploring innovation for “precise network planning and efficient network construction” so that the whole network can not only complete the conventional “flat” coverage, but also achieve three-dimensional excellent performance, which is the goal of the industry.

ZTE has independently developed an intelligent platform for wireless network planning with accuracy and precision ,ushering 5G network planning capability into a new stage.

Precise Prediction, AI Expansion, and User Experience Guarantee

Based on the AI algorithm, the ZTE AI expansion solution can predict the traffic of the whole network level, scenario level, cluster level and cell level. Traffic suppression analysis: Find the suppressed cells in time and output capacity solution to improve user experience and payload. Based on the user perception rate, the AI algorithm is used to analyze associated indicators, and the predicted traffic plus suppressed traffic is used to get the capacity expansion threshold. The precise capacity expansion is based on the capacity expansion threshold to ensure user experience. The prediction algorithm combines the advantages of SARIMA and LSTM and the LGB algorithm. Compared with traditional prediction methods such as polynomial curve fitting or Gong’s model, the accuracy is higher. After expansion, the capacity growth rate can be as high as 42%, releasing the traffic suppression, improving user perception and improving network quality.

Figure 1 Traffic suppression analysis

5G Intelligent Planning, Precision Insights, and Planning the Future

In the process of 5G network construction, operators are facing new difficulties such as innovation and complexity of 5G technologies and high network construction costs, which makes operators pay more attention to the quality of network construction. Accurate network planning is an important basis for accurate deployment of resources at the early stage of 5G network construction.

ZTE predicts 5G planning based on 4G network data (including site, PM, user number and MR) at the initial stage. 4-5G cooperates in planning in the following aspects: Multi-site hybrid site planning solution, coverage pre-evaluation, antenna weight planning and anchor planning. Based on the big data precise planning platform, ZTE implements 5G site selection, new sites, site deployment and presentation, site value sorting, The automation of planning reports and layer output improves planning efficiency and accuracy, and achieves intelligent and efficient network planning objectives.

In the stage of 5G development, the problem points are identified based on the 5G real MR(Measurement Report measurement report) data and network management data. Through all-round coverage and capacity evaluation, the adaptive deployment of macro station, micro station and indoor distribution station is realized. In 5G network planning, the particularity of 5G antennas is fully considered. By using the 3D electronic map and the principle of black and white dots, antenna weight planning is carried out accurately to create an integrated design concept of combining site planning and antenna planning. The MM weight planning model is constructed with AI technology to maximize the spectrum efficiency and network value. The average RSRP of the in-depth coverage of the building is increased by 5~10 dB.

Figure 2 4/5G RSRP evaluation based on MR data

Figure 3  Precision antenna pattern planning gain analysis

Big data +AI+ cloud computing empowers network planning

With the rise of big data, cloud computing, AI, and other innovative technologies, big data technologies enable network planning to effectively use real massive data, and enable network evaluation, planning, and design more real and valuable. Cloud computing makes the system support more powerful and resources can be configured and used more efficiently. The deep learning function of AI has a revolutionary impact on traditional network planning methods.

The future 5G network construction will be the mode of “Precision Surface Coverage Forecast, Precision Evaluation for Key Scenarios.” The improvement of site collaborative planning, outdoor weak coverage granularity aggregation, accurate antenna pattern planning, and 5G anchor point planning will make 5G innovative technologies fully play their functions and maximize the wireless coverage evaluation. The 4G MR-based 5G coverage prediction makes full use of the huge data in the existing network and supports the 5G planning and design. This function evaluates and solves the problems of high-value pain points in the network in accordance with the value dimension. ZTE’s precise network planning solution aims at the network value and achieves the precise network planning capability with the help of the big data +AI+ cloud computing. It starts with the end and helps operators build core competitiveness and a high-quality network.