Abstract
Recently, drones are being used in various work sites and the scope of use thereof is gradually expanding. However, since there is no special abnormality detection system to predict the fall of a drone, it is always exposed to the risk of a fall. Therefore, in this paper, we proposed an algorithm that predicts abnormal behavior through Long Short Term Memory (LSTM) based on the motor vibration data collected from the drone to predict the abnormal motion of the drone.