Centralized Kalman Filter for Fusion of Multiple On-Board Auxiliary Sensors with INS for Underwater Navigation

An Inertial Navigation System (INS) is capable of providing the position, velocity and attitude (PVA) data of a navigating platform based on the initial PVA and measurements of its inertial sensors. However, due to errors present in the inertial sensors, the errors in the PVA data provided by the INSs accumulate with time. In order to restrict the growth of INS error, it must be aided by other auxiliary sensors on-board the navigating platform. INS/GPS integration is a well-established way to restrict the INS error growth. However, it is not applicable for an underwater platform. For underwater applications, INS must be aided by on-board auxiliary sensors. In this paper, the integration of Electromagnetic Log (EMLog), Depth-sensor, and Gyrocompass with the INS is considered using a centralized extended Kalman filter (EKF). In a practical scenario, it is not possible to place all the sensors at the same location and each sensor measures the navigation parameter at its own placement location. Hence, a lever-arm correction is included in the proposed algorithm to transpose the sensors measurements at a common location, which is the location of INS in this case. The performance of the proposed algorithm is assessed in terms of objective criteria namely Circular Error Probability (CEP), Spherical Error Probability (SEP) and Euclidean Distance (ED). The percentage improvement in the position in terms of CEP, SEP and ED is observed to be 50.27%, 50.10% and 50.04% respectively after the integration of auxiliary sensors as compared to the position provided by the INS.