Evaluation and Comparison of Different Motion Models for Flight Navigation

Navigation performance of a flight has to be maintained with a certain level of accuracy in civilian or military operations. Global navigation satellite system (GNSS) based devices in flights act as primary source of navigation. In order to improve the accuracy and robustness of the navigation, information from other systems (e.g. IMU) are fused together. The accuracy is further enhanced when adequate motion models are used during the estimation process. In this paper, we present results of multiple motion models in association with aircraft navigation and evaluate their performances. GNSS and inertial measurement unit (IMU) data are recorded in an aerial flight for about three hours. In order to highlight the impact of different motion models, data captured is processed post flight and position estimates are computed with a linearized Kalman filter (LKF). The computed positions are then compared with a reference trajectory and errors are evaluated for all the motion models. For different flight segments, the estimated position root mean square error (RMSE) varies up to a maximum of about 4 decimeters with different motion models. Also, the magnitude of the maximum deviations in highly dynamic maneuvers is reduced by a large extent when compared with different motion models and performance improvement is about 72%.