Real-time Situation Awareness of Industrial Process based on Deep Learning at the Edge Server
Rongbin Xu, Wangxing Lin, Zhiqiang Liu, Menglong Wang, Yuanmo Lin, Ying Xie · 2020
With the popularity and application of cloud computing, edge computing, artificial intelligence and other information technologies, the virtual realization of physical world for industrial manufacturing process has become possible and played a very big role. It is urgent to find an efficient and accurate method for industrial equipment safety by situation awareness (SA) to monitor the process. In this paper, a whole life cycle of industry has been described and characterized for the operation and control of modern industrial remote control machinery. And the edge servers are employed for the operation process of tracking physical devices. Then a novel framework of graph convolutional networks (GCN) for semi-supervised classification method is proposed to find unknown patterns in huge amounts of industrial data, which can obtain real-time situation awareness for the execution of modern industry. Plentiful experiments have been conducted to show higher accuracy and stability of our framework.