Research On Cooperative Localization Method Of Multiple UAVs In GPS-Denied Environment

With the rapid development of UAV technology,it has been widely used in battlefield investigation,target attack,field rescue and so on.Different from the civil scene,this kind of complex environment is often accompanied by various obstacles and interference signals,which brings great challenges to the traditional GPS based localization methods.Without accurate positioning,the UAV can`t make correct decisions for the deployed tasks,that will affect the mission execution efficiency,and even make the UAVs drop in danger.This kind of GPS signal interference environment is called GPS-denied environment.In multi UAV application scenarios,the cooperation between UAVs can be fully utilized to improve the localization accuracy and stability of UAV cluster.In this paper,we proposed an enhanced relative localization method based on persistent excitation for Multi-UAVs.Then,we combined the positioning of other UAVs in the neighborhood for collaborative localization.At the same time,we define positioning confidence degree to adjust the proportion of UAV positioning in the collaborative process adaptively.The innovative achievements are as follows:(1)Enhanced Relative Localization Based on Persistent Excitation for Multi-UAVs in GPS-Denied Environments.In view of the rapid divergence of relative localization error in multi UAV highspeed flight scene,an optimization method based on interpolation is proposed.Firstly,the interpolation is used to fit the curve of the sampling data such as relative distance and relative velocity.Then,resampling at a higher frequency on the fitting curve is used to reduce the sampling period T at the data level.Finally,the relative localization is estimated again.Experiments show that the optimization scheme can effectively improve the accuracy of the persistent excitation relative localization method,and the faster the UAV flight speed,the better the optimization effect.In the flight scenario with the highest10m/s,the accuracy can be improved by 32%.On the other hand,with the research for the problem of error accumulation in the same speed flight scenario,two state discriminators are designed to judge whether two UAVs are flying at the same speed in real time,and the error is relatively large.If this phenomenon occurs,it is considered that the error accumulation may occur,and the relative distance is used to correct it.The experiment shows that the scheme can effectively avoid the error accumulation of relative localization.(2)Adaptive cooperative localization method for multiple UAVs based on location confidence.In order to solve the problem of UAV’s positioning error propagation in the process of cooperative positioning of UAV cluster system,a positioning confidence degree is proposed to describe the positioning accuracy of UAV,so that the UAV with accurate positioning can be more trusted in the process of collaborative positioning.Firstly,considering the operating state of each sensor that affects UAV positioning,quantifying them to adjust the positioning confidence degree.Secondly,the output of the enhanced relative localization method based on persistent excitation is used for cooperation localization,and the positioning confidence degree is used to calculate the proportion of each UAV’s own positioning in the cooperation localization process,so as to improve the positioning accuracy of the whole UAV cluster system.Finally,an Expand Kalman filtering was used to fuse the measurement data of IMU(Inertial Measurement Unit),VO(Visual Odometry),GPS and the results of cooperation localization,and output the UAV’s attitude,speed,positioning and other information for UAV control.Experiments show that this method can adaptively adjust the proportion of each UAV in the cooperation localization process,so as to improve the positioning accuracy of single UAV and even the whole cluster.(3)Design,development and evaluation of multi UAV cooperation localization system based on ROSIn view of the excellent performance of ROS simulation environment,friendly open source community support,and good compatibility with various commercial UAVs in the market,this paper uses ROS environment to design and develop multi UAV cooperation localization system.The relative localization algorithm package based on interpolation method enhancement,multi UAV cooperation localization algorithm package based on localization confidence degree and multi-sensor fusion algorithm package based on EKF are implemented and integrated into a prototype system.Finally,experiments are carried out in a multi UAV flight simulation scene built in the Gazebo three-dimensional environment.The results show that the proposed algorithm has better performance,stability and scalability than the existing algorithm.