Workshop Personnel Location Management And System Development Of Enterprise J Based On Convolutional Neural Network

China is the world’s largest industrial producer.In recent years,as the mechanization of industrial production processes has increased and personalized small batch production modes have gradually taken over,the division of personnel has become more complex,making it difficult to keep track of the distribution of personnel and their activities in large workshops in real time.In addition,with the proliferation of mechanical and automated production equipment,employees’ abnormal activities increase the frequency of mechanical automated production equipment accidents.At the same time,the current management of workshop personnel is generally in manual supervision mode,which is not only inefficient,costly and often trigger mistakes.In order to reduce the incidence of production accidents in enterprises,guarantee the production safety and personal safety of employees on duty,standardize the safety production system,and eliminate incidents such as off-duty,crosstalk and misadventure into dangerous areas,there is an urgent need to use modern information technology means to supervise the location and abnormal activities of operating personnel in production workshops.The research work conducted in this thesis is as follows:First,the background and significance of applying personnel location system in workshop production scenes are analyzed;the digital solution of personnel location management based on indoor positioning technology is proposed for the current situation of backward and inefficient way of personnel management in workshop of J enterprises.the advantages and disadvantages of various indoor positioning technologies and positioning algorithms are analyzed,and the research on workshop localization using Wi Fi-based fingerprint positioning technology is determined.Secondly,for the complex signal propagation environment of the workshop production scene,convolutional neural network with strong feature extraction and nonlinear mapping capability is introduced to improve the accuracy of personnel localization by converting Wi Fi signal strength into grayscale map as model input.In this thesis,a convolutional neural network-based indoor localization model CNNIDL is proposed,and the model is evaluated using a publicly available indoor localization performance verification dataset,and the performance is compared with classical indoor localization models such as SVM,CNNLoc and Wi Fi Net.The experimental results show that the CNNIDL neural network model has a better performance in indoor localization problems,with the localization accuracy up to 95.58% and the localization elapsed time meeting the application requirements and much lower than the localization models with the same performance.Finally,based on the efficient CNNIDL indoor localization model,a personnel localization and management system is designed and implemented for the workshop production scenario.The Java language is chosen to implement the data collection subsystem based on Android cell-phones,which realizes the functions of indoor Wi Fi location fingerprint database collection and real-time data collection.In addition,a PC-based desktop application subsystem based on Python was developed to link with the Android client to realize real-time positioning management,employee information management,risk management,intelligent attendance and abnormality identification and warning management.