Fusion of inertial navigation system with 10 axes and computer vision for UAV navigation

Calculations in this paper are based on finite difference equations. These equations are obtained by fusion of differential equations of UAV (unmanned aerial vehicle) motion and discrete measurements (carried out at a certain rate) using Kalman filter. The purpose of this paper is to fuse (using Kalman filter) an inertial navigation system with 10 axes and computerized vision-based navigation to calculate UAV position, speed and orientation. The paper includes three main sections: the first one is UAV position, speed and orientation calculation by an inertial navigation system with 6 axes, the second one is fusion of the inertial navigation system with 6 axes and sensors (altimeter, magnetometer, gravimeter) using Kalman filter (this system is defined as an inertial navigation system with 10 axes), the third one is fusion of the inertial navigation system with 10 axes and computerized vision-based navigation using Kalman filter. The paper demonstrates results of numerical simulations which prove that Kalman filter can significantly reduce the error of the inertial navigation system.