A multimodal skill learning method for improving mobile phone assembly accuracy

摘要

At present, the smart assembly of mobile phones is a challenging task. The assembly process will be affected by console vibration, illumination changes, and end self-occlusion, and it is difficult to locate directly through visual methods. In order to solve the above problems, this paper proposes a multi-modal skill learning model for improving the accuracy of mobile phone assembly. By fusing visual and tactile information, the deep learning model is used to guide the intelligent assembly of robots to improve the accuracy and success rate of mobile phone assembly. The method in this paper has carried out the experiment of flexible flat cable assembly on Redmi note 11, and can ensure the assembly error of the flexible flat cable X and Y axes within 0.3mm. This method is expected to be applied in a real assembly environment, improve the success rate of intelligent assembly and reduce the cost of mobile phone assembly.