Research On Autonomous Positioning Technology Of UAV Based On Ground Information

Unmanned Aerial Vehicle(UAV)has the advantages of flexibility,easy operation,low cost,etc.As the application requirements increase,the operating environment is increasingly diversified and complicated.In the UAV satellite localization system,the satellite signal will be seriously disturbed by the high-rise buildings,electromagnetic interference,and severe weather,which may result in failure of localization.Therefore,relying solely on satellite positioning systems can no longer meet the mission requirement of UAV.In order to improve the environmental adaptability of the airborne localization system,ensuring small size and low power consumption,research on the automatic UAV localization,which can be applied to the complex mission scenarios,has become one of the major problems to be solved.This paper mainly focuses on the UAV visual automatic localization assisted by ground information.The main research contents are as follows:(1)Three-dimensional(3D)point cloud reconstruction based on monocular sequence images.For the characteristics of UAV image data,a 3D reconstruction method based on IMU pose information is proposed.The pose information of the UAV provided by IMU is used to obtain essential matrix to realize the 3D reconstruction with the fusion of UAV navigation system data and visual information.Simulations and verifications are performed and analyzed.(2)3D point cloud registration based on conformal geometric algebra.For the large amount of 3D point cloud information,the traditional registration algorithm has the problem of inefficient and non-real time of the automatic localization.The 3D point cloud registration algorithm based on conformal geometric algebra is studied.The ICP registration method based on point line constraint is deduced,and the simulation experiment is carried out and compared with the existing algorithm.(3)The UAV automatic localization technology based on constraints of ground target.Based on the known of UAV attitudes,image coordinate of the target,geodetic coordinate and other information,the automatic localization mathematical model is constructed with the geometric constraint relationship among the ground target,image point and UAV position.The automatic localization based on single point constraint and multi-point constraint is studied with different number of ground targets,and the simulation verification and error analysis are carried out.(4)Experimental verification of autonomous positioning technology of UAV based on ground target point constraint,analyzing the influence of each measurement error on the positioning result,and proposing countermeasures according to the specificity.