摘要：In recent years, unmanned aerial vehicles have gained increasing popularity. They are employed in many applications due to its features in terms of flexibility, low cost, and small size. Due to special structure of mechanical system for the quad-rotor helicopter (typical small-sized unmanned aerial vehicle), one of the main research fields for quad-rotor helicopter is to improve its stability flight control so as to well implement its autonomous navigation. However, for the quad-rotor helicopter whose power is supplied by battery, the embedded operating system directly affects its flight performance; thus, it is of very importance to implement this goal. This article aims to put forward a high-performance software platform consisted of embedded operating system and JAVA virtual machine in the quad-rotor helicopter. Then, this article also proposes an autonomous control algorithm written by real-time JAVA application, which uses image entropy to avoid obstacles and optical flow to discover the most appropriate area to fly in while avoiding obstacles; besides, image identification with image contour detection is used for the cases of junction turn and landing. The practical experimental results show that the implemented software control platform and proposed autonomous flight application are component for indoor autonomous flight of quad-rotor helicopter.
关键词：Software control platform for mechanical system; autonomous navigation; optical flow; obstacle avoidance; image entropy; image contour
