The Key Technology Research Of Data Collection For Workshop Based On Internet Of Things

Local Internet of Things for military product manufacturing and its data management platform isplaying a significant role in the process of constructing a complete military corporation informationplatform, enhancing corporation’s manufacturing and information levels. In order to solve theproblems during collecting military industry workshop data, the paper describes the research status ofdata acquisition technology in Internet of Things, the technologies of data cleaning and data fusion.Accordingly, transparent digital workshop is constructed, which can realize intelligent manufacturingmanagement and resolve the problems that acquired data is easily to be affected by data acquisitionenvironment and devices, and that data transmission amount is too large to effectively support upperapplications. The main work of the paper includes:(1) The data acquisition system architecture of Internet of things for military productmanufacturing workshop is presented. Data acquisition mode that is both periodical and real time anda data acquisition client mode are designed. Procedures of the acquisition and processing of systemdata are provided. Key technologies involved in the system are elaborated.(2) An algorithm of RFID data cleaning based on an improved Kalman filter is proposed.Considering the characteristic of labels move frequently, sliding window technology is introducedinto Kalman filtering model. By dynamically adjusting window size, algorithm work flows of Kalmanfiltering model are improved. With the premise of data cleaning accuracy, the delay problem broughtby tag dynamic transition is resolved to enhance efficiency and accuracy of data cleaning.(3) An inner cluster data fusion algorithm based on probability transmitting is proposed.Combining data flow fusion technology with probabilistic method, a probability-based inner clusterdata fusion algorithm is implemented which considers an acceptable deviation and the correlationbetween data acquisition time and space. The algorithm can reduce network data transmission.(4) Military product workshop Internet of Things data acquisition system is accomplished. Withthe background of military product manufacturing local Internet of Things and its data managementplatform, the improved Kalman filtering RFID data cleaning algorithm and probability-based innercluster data fusion algorithm are applied into this project to accomplish requirement analysis anddesigning of data acquisition system. System network and functionality architecture and their mainwork flows are provided.