Comparison of forward Vs. feedback Kalman filter for aided inertial navigation system

An autonomous underwater vehicle (AUV) can employ the inertial navigation system (INS) to determine its position. The measurement from the INS, which utilizes the accelerometers and gyroscopes, deteriorates with time due to the accumulation of errors in the sensors. To reduce the error growth in the INS, an external aiding source such as the DGPS or the differential GLONASS can be utilized. The loosely coupled system approach is usually implemented in integrating the information from the aiding source and the INS. The two main architectures in the loosely coupled systems are the forward Kalman filter (FKF) and the direct feedback Kalman (DKF) filters. In this paper we explore the possibility of using these techniques for an aided INS in the AUV application. The advantages and the disadvantages of using the FKF and the DKF are also discussed.