Research On Robotic FPC Assembly Based On Contact State Recognition

The assembly operation of flexible printed circuit board(FPC)is an important part of the manufacturing process of 3C consumer electronics.At the present stage,the assembly of FPC is mainly realized by manual operation and auxiliary special machines,and there are problems of low automation,high labor intensity and high labor cost.In recent years,robots have been widely used in the industrial field,bringing opportunities for the development of FPC assembly operations.However,the assembly space of FPC assembly is narrow,rich in contact,and complex in relationship to the robot’s assembly strategy has raised higher requirements.Traditional industrial robots based on position control lack force feedback information and cannot describe the assembly contact process;existing industrial robots based solely on force control have insufficient ability to recognize the assembly contact state and can hardly cope with transient changes in the assembly relationship of FPC.Therefore,how to make the industrial robot accurately recognize the assembly state,guide the robot to execute the assembly action,adapt to the assembly environment of FPC,and realize the assembly task of FPC becomes the main research problem of this thesis.In this thesis,we analyze the contact characteristics of the assembly process and focus on the modeling of FPC assembly strategy and the description and recognition of the assembly contact state,which solves the problems of complex FPC assembly relationship,rich assembly contact and difficult robot operation.Excellent assembly results were achieved.The main research work and innovation points of this thesis are as follows:(1)A assembly strategy for FPC is constructed.According to the assembly process of FPC,the assembly relationship and characteristics of FPC are analyzed in detail,and the robot assembly control strategy model based on force/position hybrid control is established;the assembly process of FPC is described in stages,and the FPC assembly of robot action strategy model is proposed in each stage.The assembly contact state acquisition experiment is designed,and the alignment state and snap state,which indicate the transition of assembly stages,are analyzed in depth.(2)The LSTM-based algorithm for identifying the contact state of robotic FPC assembly is proposed.The contact state samples of the FPC are described using six-dimensional force information,and a sample calibration method based on the state signal is proposed for the snap-fit state,and a contact state recognition model based on the LSTM algorithm is constructed.The experimental results show that the method can be effectively used for the recognition of the contact state of FPC assembly.(3)An R-RNN-based online recognition method for the contact state of robotic FPC assembly is proposed.For the problems of poor security and low computing efficiency of LSTM online recognition of snap state,the online recognition framework based on R-RNN is established,and the model is optimized by using residual recurrent network unit,and the learning ability of the model is enhanced by combining with bowl-type weight function.The experimental results show that the recognition time consumption is reduced by 67.32%compared with the LSTM model.(4)Building a platform for robot FPC assembly.To validate the robot FPC assembly strategy model based on contact state recognition,a robotic FPC assembly platform based on the UR5e robotic arm is built.The findings of a wide range of assembly operation studies suggest that the strategy put forth in this study can direct the robot in order to complete the flexible circuit board assembly work.