Localization of an Autonomous Underwater Vehicle using a decentralized fusion architecture

In this paper, the position of an Autonomous Underwater Vehicle (AUV) has been estimated using the data of two estimation loops via a decentralized data fusion algorithm. Extended Kalman Filter (EKF) is used in each local loop and a decentralized Information Filter is used to fuse the data obtained from the other loop. The sensors used in the loops are: loop 1) Doppler Velocity Log (DVL), rate gyros as internal sensors, and a pressure sensor and compass as the external sensor, loop 2) accelerometer, two inclinometer and Z-axis free gyro as internal sensors and echo sounder as the external sensor. AUV can be localized using each of the two estimation loops, but the decentralized architecture is more robust and leaves a degree of redundancy for checking possible faults of sensors and/or local estimation algorithms. The results show that despite the limitations in choices and arrangements of the sensors, the two local loops perform appropriately and the fusion of the estimates of the local loops improves the robustness of the estimates. At the end, the proposed decentralized architecture has been compared with a centralized algorithm and its advantages are pointed out.