Research On Integrated Navigation Of UAV Based On Inertial/GPS/Optical Flow

In recent years,whether civilian or military,the use of drones has been increasing.In the future,drones will occupy an important position.The drone's navigation accuracy is the key to its mission.In today's complex environment,drones need to face more and more situations.When performing tasks,drones need to face complex problems,such as small flying space,some sensing devices are destroyed by the enemy,and Ensure the accuracy of long navigation.The traditional drone navigation mainly relies on inertial devices or satellite navigation methods.The inertial navigation alone will have the accumulated error of the device,and the satellite navigation is vulnerable to the destruction of the enemy during the war.Based on the research of traditional navigation,this paper combines the navigation methods.Under the traditional navigation method,computer vision data based on optical flow is introduced to integrate data for multiple navigation methods.Finally,the UAV hardware platform is used for verification.The main work of the dissertation includes:1.Based on the design of drone integrated navigation scheme,this dissertation first studies the principles of traditional inertial navigation and GPS satellite navigation,and establishes relevant simulations.Based on this,a strapdown inertial navigation method based on quaternion method is proposed.The scheme,combined with GPS,proposes a coupled navigation method based on Kalman filter algorithm.Verify the feasibility of traditional integrated navigation methods in navigation tasks.2.The concept of how to ensure the navigation accuracy of UAV when the traditional integrated navigation method is disturbed is introduced.The concept of optical flow navigation in computer vision is introduced.The basic principle of optical flow navigation is discussed and an ORB-based feature point is proposed.LK Pyramid tracking algorithm,based on this algorithm design experiments to verify.After verifying the relevant optical flow algorithm,the 3D UAV is combined with 2D image points to study the relationship between the image flow field and the drone sports field,and an optical flow sensor based solution is proposed.Machine navigation method,using relevant simulation experiments to verify.3.For the optical flow image data of drones,due to the influence of the environment,the image data obtained is biased.This paper proposes a new form based on inertial/optical flow integrated navigation,using tightly coupled Kalman filter algorithm.The traditional integrated navigation method was supplemented and the integrated navigation method of the drone under special mission scenarios was improved.After researching the theoretical algorithm,it is necessary to design a corresponding hardware platform for verification.This paper designed a four-axis UAV hardware platform based on PIXHAWK flight control and related sensors for this problem,and proposed a four-axis system based on PID control model.The drone control scheme,combined with the application of the optical flow sensor in the actual scene,uses different navigation methods to perform the verification experiment respectively.The experimental results show that the scheme studied in this paper has strong feasibility.