Research On Vision Assisted Control Of Quadrotor Aircraft

With the development of micro-aircraft,the quadrotor UAV has gradually entered the field of view of the general public.Due to the flexible control of the quadrotor,simple structure,small size and many other advantages,The application of the quadrotor in the civil field of aerial photography,rescue,security and other fields is increasing,and the potential engineering application value reflects the broad market prospects.Although the application value of the quadrotor is large,there are also problems.For example,the control efficiency of the quadrotor in the mountainous areas without GPS and the small indoors is greatly reduced.For these limitations,this paper independently builds the software and hardware platform of the quadrotor UAV,designs the image optical flow module for the drift problem without the positioning system.At the same time,the attitude calculation,attitude control,position control problem are studied.The specific research contents of this paper are as follows:1.By analyzing the flight principle,motion mode and structure of the quadrotor UAV,establish a coordinate system describing the motion and force analysis.In addition,the motion of the quadrotor is divided into two modes: linear motion and angular motion.Finally,the dynamic model of the quadrotor is established under simplified conditions.2.An image optical flow velocity measurement system is designed.Aiming at the problem that the quadrotor is difficult to control when there is no positioning system such as GPS in the interior,the optical flow velocity measurement system based on SAD block matching algorithm is designed,and the error of the two-dimensional image during the rotation of the quadrotor is compensated.A communication protocol between the image device and the quadrotor is designed.After experimental testing,the system can effectively estimate the horizontal drift speed of the quadrotor.3.The actual control of the quadrotor was studied.Firstly,a pose fusion algorithm based on vector cross product PID is proposed for the drift error of gyroscope,which realizes the accurate solution of the attitude angle of the quadrotor.Secondly,the attitude control algorithm and height control algorithm based on cascade PID are designed to realize the attitude control and height control of the quadrotor.Then,under the airborne image optical flow module,a position control algorithm based on cascade PID is designed to realize the hovering control of the quadrotor.4.Finally,for the horizontal drift of the quadrotor,the Kalman filter is used to estimate and eliminate the error that is difficult to observe by the accelerometer.On this basis,the optical flow velocity is combined to obtain a more accurate horizontal drift velocity,which improves the position control accuracy of the quadrotor.After the actual flight test,the attitude control,the height control,the position control of the added optical flow,and the hovering position control through the Kalman filter fusion all achieved the desired effect.