A resilient power fingerprinting selection mechanism of device load recognition for trusted industrial internet of things

In order to monitor the stability of industrial systems, engineers installed diversified sensors in systems, and used communication devices to transfer the sensed data to the cloud platform for real-time monitoring and event detection. Furthermore, as industry demand for power grows, the scale and quantity of power systems gradually increase, and the original network data transmission architecture cannot bear such large-scale communication, especially the communication bandwidth tolerance isn't allowed for trusted industrial Internet of things. Therefore, this trusted transmission problem will be one of challenges of the industrial Internet of things. In the application of device load recognition, how to create power fingerprinting recognition sample data, reduce the cloud platform computation complexity and the transmission quantity of sensed data without losing detection accuracy are the subjects of this study. Therefore, this study proposes a resilient section selection mechanism of power fingerprinting applied to device load recognition, in order to determine the transmission time and select the power fingerprinting section to be resiliently transferred, and replace the cycle-fixed full power fingerprinting data transfer for trusted industrial Internet of things. According to the experimental results, in the case of multi-load, the power fingerprinting of the first 25% section have the maximum recognition of 87.5%.