The world's first example of Ericsson providing So

  • Detail

The first in the world! Ericsson provides Softbank with network design based on machine learning

recently, Softbank, a leading mobile operator in Japan, adopted Ericsson's innovative solutions to design wireless access networks based on machine intelligence. The service groups the cells and considers the overlapping coverage of cells and the feasibility of using carrier aggregation between cells. It can reduce operating expenses and improve network performance. Compared with the traditional network design method, this scheme can shorten the lead time by 42% and simplify the installation by 0%

Softbank adopted Ericsson's services in the complex and dense urban agglomeration in the East China Sea, and the effect exceeded our expectations. We are working closely with Ericsson and plan to deploy the same services in other regions

ryo manda

wireless technology manager of Donghai Network Technology Department of Softbank

this method is based on a comprehensive analysis of the actual wireless network environment, and covers many aspects, such as cellular overlap coverage, signal strength and receive diversity. Considering the possible connections between cells and the needs of network evolution, wireless network design needs powerful computing power and first-class machine learning technology

in order to complete this complex and challenging task, Ericsson adopted the cutting-edge design concept of patented network graph machine learning algorithm (community discovery)

despite the challenges, Softbank successfully used Ericsson's services to automate the wireless access design process, applied big data to the cluster composed of 2000 wireless cells, and analyzed the data to obtain optimal configuration

peter laurin

head of Ericsson Telecom Management Service Department

machine learning has great application potential in the telecom industry, and Ericsson has invested a lot of resources in this technology. We are very pleased to see these new methods successfully applied to the network of Softbank. Top operators in other regions have also shown strong demand for the deployment of this solution

Ericsson combines a wide range of wireless network capabilities with the latest machine learning progress, providing customers with key differentiation factors on the road of automation. The AI accelerator laboratories operated by Ericsson network design and optimization department in Japan and Sweden are developing these use cases to support Ericsson's global strategy of introducing machine intelligence into different well-developed enterprise fields

Ericsson's centralized and flexible wireless access network design method will be sold in the form of services and support LTE networks

the annual world telecommunication day

the theme of this year is exactly:

promoting the proper use of artificial intelligence (AI) for the benefit of mankind

since 2007, the Research Laboratory of Ericsson, which loves measuring tools with an accuracy of more than 0.1 mm, has been exploring the automation technology driven by machine intelligence, and has obtained hundreds of relevant patents. Ericsson has made active layout and aims to lead the automation revolution in the communication industry

intelligent network operation

in the past, all networks around the world were manually managed by humans. Thousands of alarms appear on the screen of the network operation center every day. Engineers need to monitor these alarms, create fault lists and solve these problems. In today's network operation center, only 5% of fault lists are created manually, and the rest are created automatically. Before that, the engineer must browse all the fault sheets before he can evaluate what corrective tasks need to be taken and make corresponding descriptions in the work order. In 2017, we handled 51million fault tickets. It can be imagined how difficult it would be for engineers to manually check these faults and assign work orders

in addition, we have established automation rules in the network. The system will process information in real time and accurately execute the designed automation task sequence when a specific activity is triggered. The automation engine sends about onemillion instructions to the network every day and automatically handles 70% of the alerts

The concept of intelligent network operation also emerged. Using the available data in the network and various other forms of external data such as weather conditions, combined with machine intelligence technology, makes the network operation more intelligent, and helps enterprises deal with complex communication systems more efficiently. Using machine intelligence technologies such as advanced analysis, deep learning, machine learning and artificial intelligence, we can create forecasts and models that human beings cannot create independently, and promote automation to a new level. Intelligent automation can predict events before they occur, and take preventive or corrective measures to avoid accidents

Ericsson will continue to invest in machine learning and machine reasoning, and make full use of the inherent opportunities of emerging network technologies. Ericsson has unique advantages in global influence, industry insight, professional knowledge and machine intelligence concept, and is committed to further opening the growth space to make intelligent network operation a reality

Copyright © 2011 JIN SHI