Improvement of sensor data used in positioning of unmanned aerial vehicles using filter techniques

With the advancement of technology in the world, unmanned aerial vehicles (UAVs) are used to serve various fields such as defense industry, transportation, reconnaissance, agricultural spraying and firefighting. In order to successfully utilize UAVs in these areas, it is necessary to determine the positions of UAVs and to make improvements on their existing positions. Global Navigation Satellite Systems (GNSS) based systems are used for determining the position and in case of interruption of GNSS signals, Inertial Navigation Systems (INS) are used. The obtained position information is passed through various filtering algorithms and data improvement operations are performed. In this context, the most commonly used filtering algorithms are Kalman and Extended Kalman Filter algorithms. In this study, faulty GPS and INS data used in the localization of unmanned aerial vehicles have been tried to be improved by using various filtering techniques and these techniques have been compared. For the linearly acting UAV, the faulty GPS measurement data generated in the simulation environment was tried to be corrected by means of Kalman and Information Filters. According to the results, Kalman Filter performance was superior to Information Filter performance. For nonlinearly acting UAV, GPS and INS measurement data produced in simulation environment were tried to be improved by using Extended Kalman Filter (EKF) and Extended Information Filter (EIF). According to the results obtained, although the EKF gives more successful results than the EIF in the improvement of the GPS measurement data, it has been found that the EIF gives better results than the EKF in the correction of the position information provided by INS.