Control method of shaft and hole mating based on convolution neural network in assembly building prefabricated components

3rd China-Romania Science and Technology Seminar

自然科学;工业技术

Zhang, K.^1 ; Tong, S.^1 ; Zhao, J.^1 ; Wang, Q.^1

College of Mechanical Engineering, Shenyang Jianzhu University, China^1

关键词: Building construction;    Convolution neural network;    Convolutional neural network;    Multi label classification;    Positioning control;    Prediction accuracy;    Prefabricated components;    Search Algorithms;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/399/1/012061/pdf
DOI  :  10.1088/1757-899X/399/1/012061

来源: IOP

【 摘 要 】

In order to further improve the automation degree of assembly building construction, realize prefabricated component intelligentized positioning, this paper applies the convolutional neural network in positioning control in the field of prefabricated component of assembly building, and puts forward a probability search into target region to establish a new method for loss equation algorithm, the method based on multi label classification loss equation modelling with probability search algorithm, in the training process to continuously improve the positive incentive probability score. The simulation results show that the designed convolution neural network can achieve a correct prediction of the assembly location control instruction, and the proposed method can improve the prediction accuracy and convergence of the network model.