Construction of gravity gradient referenced inertial navigation system and its performance analysis

To overcome the weakness of the global navigation satellite system (GNSS), gravity gradient referenced inertial navigation system (GGINS) based on the extended Kalman filter (EKF) has been developed. The performance of GGINS has been verified considering DB/sensor errors, flight altitude, DB resolution, initial errors as well as update rates based on the simulation tests. As expected, GGINS generates better results when DB and sensor errors are smaller and flight altitude is lower. When compared the performance of the GGINS with the well-known terrain referenced navigation (TRN), DB and sensor errors should be 0.1Eö and 0.01Eö, respectively to obtain better or similar performance. Additionally, navigation performance worsens when DB resolution is lower and the initial error is supposed. However, effect of initial errors has been compensated and similar navigation results have been obtained after convergence. In case no initial errors are supposed, GGINS based on 10 epochs update rate generated better results. However the performance worsens as update rate is getting larger. Especially, it should be updated and compensated every epoch when initial errors exist. Therefore, it was difficult to determine certain update rate to be general for all simulation trajectories. To construct more effective algorithm, two update condition using relation between measurements and DB/sensor errors, and the selection of components were tested. Some trajectories showed better performance owing to specific update conditions but it was hard to be generalized. Therefore, additional study to construct robust algorithm is necessary in the future.