New network technologies are constantly iterating, putting forward higher requirements for autonomous operation and maintenance of communication networks, and the requirements for communication services in various sectors of society have far exceeded the scope of traditional services. By implementing the Smart 5G business philosophy of being smarter, more agile and greener，our Catalyst project will solve the needs of the transformation from traditional communication services to new information services, improve the supply capacity of information services, promote the coordinated development of the upstream and downstream of the 5G industry chain, and help realize the value goal of the digital intelligence transformation of the whole society. In promoting the project forward，there are three major challenges： Challenge 1, How to deeply integrate AI technology with 5G new technology, and embed AI and data intelligence into all aspects of 5G operation and network operation and maintenance is still the focus of current operators. Aiming at reducing enterprise operation and maintenance costs, improving operational efficiency, and upgrading network intelligence through self-configuration, self-optimization, and self- healing. Challenge 2, How to collect and sort out user perception data by intelligent means, predict business quality and user experience changes in advance, and finally achieve network supply capacity based on user experience needs, and consolidate customer understanding and heart-level service feelings is still a major topic. Challenge 3, How to achieve hierarchical management of network self-intelligence capabilities, provide differentiated network information services according to the needs of various industries in society, incubate more business scenarios, help various industries improve service capabilities, green management, and help realize the transformation of social digital intelligence is the biggest challenge we are currently facing. The project proposes targeted technical solutions for 5G network operation and maintenance nodes to form the technical support foundation, solve the above challenges, and fully support the application needs of users and various business scenarios; For public network users, it mainly predicts user distribution, user demand and user behavior, and provides accurate and agile user perception guarantee;For business scenarios，it provides differentiated network services. The project has catalyzed 8 use cases： UC1 Autointelligent network：AI enables the 5G antenna beam autonomous optimization Pain points：Mobile network users are constantly moving and changing positions, and fixed grid coverage cannot meet the needs of mobile users, and manual operation and maintenance costs are high and timeliness is low. Solutions：The project innovates the spatio-temporal recurrent neural network AI algorithm, using the user's 7*24 hours location distribution information to achieve accurate prediction of user distribution; Using machine vision algorithm, the multi-cell joint optimization model is established to quickly obtain the optimal configuration of antenna beam parameters, the whole process is intelligent. Benefits：The system has been widely used in large supermarkets, universities, office buildings and other scenarios.The project achieved an 11% increase in user perception, an 8.4% increase in 5G business, a 5.2% decrease in 5G single station maintenance costs, and advanced the autonomous level in the beam scenario from L3 to L4. For the first time in our country, the network moving is intelligent adjoint with user moving. UC2 BS anomaly： compensation mechanism for 5G network abnormal events Pain points：At present, most wireless site faults need to be handled onsite, and the processing period is long. During the troubleshooting, the network experience of users in the surrounding areas is seriously affected. Solutions：We periodically scan and capture site faults to quickly locate the fault area.In view of the possible "pay attention to this and lose the other" problem in the compensation process, the innovation of "iterative adjustment convergence algorithm" confirms the compensation adjustment range, and immediately calls the surrounding area to cover and supplement once the site breaks down, ensuring the overall stability of the network while guaranteeing the compensation effect. Benefits:The project has been applied in the vertical industries such as smart park, smart police, smart transportation and other 12 business scenarios.This project can recover 92% of business loss and reduce the number of user complaints by 36%. UC3 Environmental anomaly: resuming mechanism for environmental anomalies response Pain points：The ability to catch anomalies in the network environment is poor, and there is a lack of effective and timely network solutions. Solutions：The system combines AI algorithm with big data model to train proprietary performance and user perception data to form a large communication model, identify network emergencies and user behavior changes, search for optimization through ant colony algorithm, formulate optimal adjustment schemes for weight, interoperability, scheduling and power coordination, establishes network burst self-optimal response guarantee scheme, and improves network response agility. Benefits：Abnormal event processing time is reduced by 60%, customer response agility is increased by 45% and abnormal event handling operation and maintenance costs are reduced by 32%. UC4 Individual user：Intelligent identification and scheduling Pain points：Diverse service types and multiple network data paths lead to complex user behavior and perception problem location, which makes it difficult to provide accurate network services. Solutions：The project takes XDR CDRS, 28 kinds of KQI indicators of service size, satisfaction score data, and site parameters as input sources. KNN algorithm is used to classify the features of cell services in different areas, and use neural network training models to establish comprehensive perception scoring systems under different scheduling schemes to output optimal scheduling schemes. Benefits：The system has been deployed in residential areas, office buildings, supermarkets, universities, scenic spots and other scenes.After deployment, network KQI and satisfaction have significantly improved, and the complaint rate has significantly decreased. UC5 Business user：Enterprise application of electricity energy saving Smart 5G project provides differentiated network service solutions according to the application requirements of scenarios. Pain points： Chizhou ultra-high voltage (UHV) as the main artery of west-east power transmission, because of the complex mountainous area, the line inspection is difficult and the cost is high. Solutions：China Mobile Anhui has partnered with ZTE to develop an end-to-end 5G technology solution,embed smart 5G technology solutions, establish DRA business identification capabilities.Through the dedicated UPF and network slice, through the collection of fire, ice and other hidden danger information, to build a panoramic intelligent control operation and maintenance platform, to achieve the "dual network integration",through the 5G drone and artificial intelligence analysis, the development of the mobile end to replace the fixed end, with machines to replace people intelligent routing inspection system. Benefits：The cost of manual inspection is reduced by 2.2 million yuan per year. UC6 Business user: Enterprise network application of factory energy saving Pain points：How to assess the needs of smart 5G factories to maximize cost reduction and efficiency. Solutions：Through the demand assessment of the intelligent 5G factory, the intelligent 5G project strengthens the advantages of 5G high bandwidth and low latency, prompting JAC to connect more than 3,60 production equipment, and the project combines AI algorithms for intelligent control, real-time collection of energy consumption information, and finally realizes robot inspection and intelligent logistics. Benefits：The solution reduces manual intervention, improves equipment utilization, and reduces carbon emissions by 281 tons per year. UC7 Business user: XR smart campus Pain points：At present, there are some problems in education, such as learning resources limited by regions, unable to meet the different needs of students, management difficulties, and security risks. Solutions：Based on the business identification database, this project intelligently identifies XR business, independently learns XR business characteristics, and establishes an XR evaluation system. Combine digital twins to improve evaluation efficiency and accuracy, beat intelligent scheduling to improve efficiency, multi-user anti-resonance, and fully guarantee teaching experience.The system provides 5G network with low delay, high speed and high reliability to meet XR business requirements for smart campus. Benefits：It has been widely used in University of Science and Technology of China, Huangshan University, Fuyang City Education Bureau, etc.Learning resources increased by 35%，learning experience improved by 60%， safety factor increased by 80%. UC8 Business user: Intelligent track inspection Pain points：At present, there are some problems in track inspection such as inaccurate data collection, low efficiency of data collection and low efficiency of manual inspection. Solutions：According to the characteristics of the track business scenario, the project provides powerful data processing capabilities, resets the QCI-level intelligent scheduling priority and MIMO self-optimization capability of 5G network, provides the best information interaction experience for the entire inspection system, forms real-time interaction between the front and back end, and finally realizes the human-computer interaction capability. Benefits：It realizes the robot to replace the manual on-site inspection.The operation and maintenance efficiency of Hefei rail transit inspection project is increased by 65%, and the maintenance cost is saved by 30%.