Linear motor primitive generation method based on observation spots

In recent years, research has been actively conducted on motor primitive generation techniques for autonomous flights of UAV using demonstration-based learning. Previous studies on the generation of motor primitive using demonstration-based learning aimed to replicate the movements observed. However, in order to achieve the various purposes of UAV, it is necessary to have motor primitive generation method that includes flight environment necessary for its operation and dynamically generated additional UAV metadata. The purpose of this study is to propose motor primitive generation process, in which the surveillance UAV dynamically generates the waypoints necessary for flight based on the path recorded by the pilot.