Paper Title
An LSTM Model to Predict The Subjective Preference for Robot Trajectories in a Context of Human-Robot Collaboration

Abstract
In the process of human-machine cooperation, the acceptance of the path and submission method of the robot arm can be improved, and the robot arm can meet the expected effect of the human when it moves, which can improve the efficiency and fluency of the human operation, and meet the human The posture of psychological expectations can also improve safety and reduce fatigue during the operation. This study will focus on the subject's path planning and submission postures randomly generated in different locations as factors, and in the experiment, use different submission postures and random movement paths in different gripping objects as the entire submission process The subject provides subjective data on the movement path or posture of the robot arm after the experiment. After data collection, the deep learning calculation is used to extract each part of the path to extract the most suitable movement trajectory and delivery method for different delivery objects and positions. In addition, the objective data is analyzed, and the range and posture of the moving path are effectively analyzed and summarized. Keywords - Hand-Over Task, Human-Robot Collaboration, LSTM, Predictive Models