UAV Landing Information Prediction And Attitude Control

UAV landing Information's Extraction,prediction and attitude control are the current hot topic. This topic is of great significance for the consolidation of our national defense, the strengthening of our economic strength. The research of the paper roots from the National Natural Science foundation-funded projects - "The real-time UAV's landing guidance information's extraction and conversion." The main program is using airborne camera equipments to take photo of the runway scene. Critical information for the UAV's landing was extracted out after the image processing to the sequence of the runway. Through this critical information, the aircraft's flying condition is extracted out and the following-up vehicle and camera's states are also tried to be predicted and controlled. In this paper, the sequences of runway scenes images were processed to extract out the perspective invariant features. Of course, these features can be point, line, graphics with perspective invariance, and the centre moments after comprehension. After access to these perspective invariant features, the method to the UAV and the camera's gesture information was gained by looking up relevant literature.After the acquisition to the camera gesture information, this paper presents a filtering prediction basing on the credibility of the runway tracking information to replace the traditional kalman filtering process and successfully estimated the UAV's information. After simulation experiments, the method was confirmed with high efficiency, good tracking ability.After the state estimation, this paper gave launch conditions of the UAV's control, according to the current international advanced level landing. In the actual control process, this article discussed the UAV control curve and gave a relatively smooth landing adjustment program. Additionally, the article discussed the UAV's head control method based on image .At last, this article pointed out the work direction in the future.