Research On Emotion Recognition Based On Skin Electric And Safety Early Warning Model For Hoisting Workers

The development of assembled construction is a major force in promoting the transformation and upgrading of China’s construction industry,and it has unique advantages in improving engineering construction efficiency,enhancing project quality and safety,energy saving and environmental protection.However,as the production mode has shifted from on-site construction to factory production of components and on-site assembly,the safety of lifting operations has become an issue that requires extra attention.This paper takes the unsafe behavior of hoisting workers as the research object and applies machine learning for safety early warning of assembly building hoisting operation,aiming to identify and monitor the emotional condition of hoisting workers by monitoring the electrical skin signal of workers and using KNN classification algorithm,and construct a safety early warning model of hoisting workers based on electrical skin emotion identification by analyzing the tendency of unsafe behavior of workers under different emotions,through which realize the early warning of workers’ unsafe behaviors when they are in different emotions.The specific study contents and results are as follows:(1)The relationship between assembly hoisting workers’ emotions and unsafe behaviors was analyzed.Through the questionnaire survey,the common negative emotions of hoisting workers are derived;after that,based on the statistical data of crane accidents,the main causes of crane accidents are analyzed,and it is found that the unsafe behaviors of hoisting workers are one of the main factors leading to the occurrence of crane safety accidents;then based on the data of safety accidents,the common unsafe behaviors of workers are identified,and through the distribution of questionnaires,the study composes the workers under different emotions The unsafe behavior tendency of workers in different emotions.(2)Emotion-evoking experiments were designed and relevant electrodermal data collection was completed.The material that can best induce the target emotion of the hoisting workers was firstly screened through the experiment;on this basis,the experiment was designed to induce the emotion of the workers,and the workers were allowed to carry out the simulation operation of the tower hoisting operation under the target emotion.By wearing a wireless sensing skin electric collection device,the skin electric signals of hoisting workers performing hoisting operation simulation under different emotions were collected,and the signals were pre-processed to extract a total of 34 features in the time-frequency domain;however,since the skin electric signal features and emotions were not completely correlated,the non-parametric test method was applied to further screen the features,and finally the effective ones with significant differences in performance under different emotions were extracted 24 features were extracted.(3)A safety warning model for hoisting workers based on skin electrical emotion recognition was constructed.The effective features of the skin electrical signal are used as input variables,and three classification algorithms,KNN,SVM and plain Bayes,are applied to build the emotion classification model.By combining the emotion classification recognition model based on skin electrical signals with the research results of the relationship between workers’ emotions and unsafe behaviors studied in the previous chapters,a three-way relationship framework of "skin electrical signals-emotions-unsafe behaviors" is built,and a safety early warning model based on skin electrical emotion recognition for hoisting workers is constructed.Therefore,this study realizes the monitoring of the emotional state of hoisting workers by collecting skin electrical signals,and then inferring the unsafe behaviors of workers,thus providing a basis for managers to take effective and targeted emergency management initiatives.The application of this result can reduce the incidence of crane safety accidents in assembled buildings to a certain extent and improve the safety management level of assembled building construction.