State Estimation of UAV using Extended Kalman Filter

Abstract: In unmanned systems an autopilot controls the outputs of the vehicle withouthuman interference. All decisions made by the autopilot will depend on estimatesdelivered by an Inertial Navigation System, INS. For the autopilot to take correctdecisions it must rely on correct estimates of its orientation, position and velocity.Hence, higher performance of the autopilot can be achieved by improving its INS.Instrument Control Sweden AB has an autopilot developed for fixed wing aircraft.The focus of this thesis has been on investigating the potential benefits of usingExtended Kalman filters for estimating information required by the control systemin the autopilot. The Extended Kalman filter is used to fuse sensor measurementsfrom accelerometers, magnetometers, gyroscopes, GPS and pitot tubes. The filterwill be compared to the current Attitude and Heading Reference System, AHRS, tosee if better results can be achieved by utilizing sensor fusion.