Research On The Technique Of Optical Flow/INS Navigation System In Multi-rotor

Because the multirotor unmanned aerial vehicle(UAV) has the advantages of VTOL(vertical takeoff and landing), structure simple and some other advantages, in recent years, it has become quite a hot research field. On usually,the multirotor UAV navigation system has a heavy reliance on the GPS, which in some cases GPS signal is disturbed or lost, it will not be able to provide effective location information. This is one of the common causes that currently rotorcraft flight accidents occur. The optical flow navigation system inspired by the vision-based navigation of insects has the characteristics of small size, low cost and high precision. It is considered as one of the ideal solutions to complete autonomous navigation of multirotor UAV's without GPS. In this paper, based on the optical flow navigation algorithm, the research and implementation of the tightly-coupled optical flow / INS navigation algorithm for multirotor UAV is carried out.Firstly, this paper studies the optical flow algorithm. Then, in view of the characteristics of multirotor UAV's flight environment, the optical flow algorithm and its error compensation algorithm are studied in large scale motion, and a LK Pyramid Optical Flow algorithm based on ORB features is proposed. Finally, the performance of the proposed algorithm is verified by static and dynamic experiments, which provides technical foundation for multirotor UAV navigation using the optical flow information.In order to provide autonomous navigation information for multirotor UAV using the optical flow information, this paper studies the relationship between the image optical flow field and the motion fields of multirotor UAV. Taking into account that the accuracy of the height information will affect the accuracy of optical flow velocity measurement, a new information fusion scheme is proposed aiming at the altitude channel. After analysis the motion pattern, designed a multiple sensors assistant optical flow autonomous navigation algorithm carrying out indoor and outdoor experiments. The experimental results show that the scheme can provide more accurate navigation information for multirotor UAV.In order to improve the reliability of the optical-flow-based autonomous navigation algorithm, this paper models and analyzes the error of each sensor in integrated navigation system, and studies the loosely-coupled optical flow / INS filtering equation based on the Kalman filter. Through the simulation experiments, the result that navigation information from loosely-coupled integrated scheme is more reliable compared to the optical flow scheme is validated.On this basis, in order to make in-depth fusion between the optical flow information and inertial navigation information, this paper has analyzed the relationship between velocity information of the inertial navigation system with image optical flow velocity, and has established tightly-coupled optical flow / INS filtering equation. The simulation experiment results show that tightly-coupled scheme is more reliable compared to the loosely-coupled scheme.At the end of this paper, an experimental verification platform based on the quad-rotor aircraft has been built, and the optical flow / inertial integrated navigation system has been constructed. In order to observe the flight status of multirotor UAV in real time, a UAV autonomous flight ground station based on LabVIEW has been developed. At the same time, the ground station has the functions of three-dimensional simulation of flight environment, sensor fault detection and so on. Finally, the system is used to verify the reliability of the tightly-coupled optical flow / INS navigation algorithm based on the actual flight tests. The experimental results show that the schem proposed in this paper could be used as one of the alternative schems for autonomous navigation of multirotor UAV without GPS.