Research On Development And Analysis Method Of Internet Of Things Data Acquisition Based On Fog Computing

With the continuous development of the Internet of Things technology,the IoT environment awareness and data processing technology have received extensive attention from researchers.Although existing environmental awareness and data processing technologies can efficiently complete data collection,processing,and analysis,in the context of the growing number of IoT access devices,IoT applications built on related technologies will face challenges in terms of bandwidth,data processing capacity,system energy consumption and other aspects.In response to the above problems,this thesis studies the method of data collection and analysis based on fog computing.The main work and innovations of this thesis are as follows:(1)Aiming at the high energy consumption of the existing data acquisition and analysis system,a data acquisition and analysis system based on fog computiong and an adaptive data acquisition strategy are designed,and the system is implemented based on the proposed strategy.The test in the actual scenario shows that the data acquisition system based on this strategy can perceive the change of environment more quickly and has higher energy utilization than the existing system.(2)Aiming at the problem of deploying real-time data processing application of IoT in resourcelimited devices,the offline and real-time hybrid data analysis architecture based on fog computing and the data compression and anomaly detection methods based on incremental principal component analysis are proposed.Aiming at the problem of model accuracy decreasing when the method processes real-time data,a regular-SPE principal component model updating strategy is proposed.The test results on the public dataset and the actual hardware equipment show that the proposed architecture and method can realize real-time data compression and abnormal data detection,improve the real-time performance of data processing applications and lower the overall resource requirements of the system.The proposed model update strategy can update the model on demand with low resource consumption,further reduce the overall resource consumption of the system,and make the method maintain high accuracy in a variety of abnormal data distribution scenarios.