A Practical Algorithm for Deep-sea AUV Vertical Channel Navigation Based on IPS/INS/CTD

In the field of Autonomous Underwater Vehicle (AUV) navigation, researchers have mainly been focusing on the establishment of the vehicle horizontal channel, and little research on the vertical channel navigation. However, vertical channel navigation is critical for AUV deep-sea missions. AUV is commonly equipped with Intelligent Pressure Sensor (IPS) and Inertial Navigation System (INS) to obtain vertical channel kinematic parameters. Nevertheless, the accuracy of straightforward conversion from pressure to depth is challenged due to the oceanic dynamic environment. Furthermore, velocity estimation error in the vertical channel of INS will exponentially diverge during the long time diving and purely inertial navigation. Vertical channel instability of AUV not only is a threat to the safe operation and recovery of AUV but also greatly degrades the quality of gathered data for the deep-sea missions.To circumvent the aforementioned issue, we proposed the Globicephala algorithm to enable the accurate estimation of the AUV vertical position and velocity in our research. We first use the UNESCO 1983 algorithm and the vertical profiles of seawater temperature and salinity, measured by the shipboard SBE 911plus CTD(Conductivity-Temperature-Depth sensor), to construct an accurate pressure-to-depth conversion database. And then we use the Adaptive Kalman Filter (AKF) to fuse the IPS, INS, and CTD data to generate an optimal solution of vertical position and velocity. A series of actual sea trials suggested that the proposed Globicephala algorithm can provide reliable and accurate vertical channel estimation. It is a great improvement for the AUV vertical channel navigation.The Globicephala algorithm fills the lacuna of AUV vertical channel navigation to some degree. It can also be used for other types of underwater vehicles, such as the Remotely operated underwater vehicle (ROV) and the Human Occupied Vehicle (HOV).