Intelligent monitoring solution based on KubeEdge

Challenge​

It is an important demand for the manufactory of Raisecom Technology to ensure the industrial production safety. Traditional workers' production safety was detected manually, which was slow and inefficient. The situation that workers did not obey the safety requirements still happened, and it could be ignored sometimes, which could generate great safety risks and affect the production efficiency of the factory.

Solution​

An industrial intelligent monitoring application with AI algorithms was developed to replace the manual method. An intelligent application alone was not enough and new problems arose such as the deployment and management of the intelligent edge application and the collaboration between training on the cloud and reasoning on the edge, which could become a bottleneck for the largescale application of the solution in the industrial production environment.

China Telecom Research Institute used KubeEdge as an important part of the implementation of the intelligent monitoring solution to help Raisecom Technology to solve the problem. Architect Xiaohou Shi from China Telecom Research Institute completed the design of this solution. In this case, the safety status of factory workers was monitored by the industrial vision application in real time with the deep learning algorithm. KubeEdge was introduced as an edge computing platform for the management of the edge devices and the running environment of the intelligent monitoring application. The monitoring model could be trained on the cloud and deployed to the edge nodes for reasoning execution automatically via KubeEdge, which could improve the efficiency of the operation and reduce the cost of the maintenance.

Impact​

In this application scenario, KubeEdge completed the unified management of edge applications. KubeEdge could also make full use of the advantages of the collaboration of the cloud and edge. With the help of KubeEdge as the edge computing platform, the monitoring on safety of the manufactory with AI was completed effectively, which reduced the occurrence of safety accidents and improved the production efficiency of the manufactory.

Based on this successful case, more deep learning algorithm will be deployed on KubeEdge to handle problems on edge computing. More cooperation about scenario-faced industrial intelligent application with KubeEdge will be carried out in the future.