Board-to-Board connector mating using data-driven approach

Abstract

Force measurement and control for the automatic process are crucial in automation, especially in the insertion (mating) task. This fragile task is needs to be automated for safety and economical purposes. One small mistake and misjudgement by operators could damage the fragile component, and also cause the company material loss. In this paper, the mating process is implemented by an articulated robot with a force sensor mounted on it. We propose a data-driven approach for the procedure to automate the mating process of the slimstack Board-to-Board (BtB) insertion process. The force data is recorded and encoded to a recurrence 2D plot. Then the 2D image is used to predict the position and alignment of the male and female Board-toBoard connector. By using the encoding approach, the system can classify each corresponding force based on its status of BtB insertion and provide a safety procedure in the insertion process. The proposed model is compared with the efficient time series LSTM model.