A method for on-line water current velocity estimation using low-cost autonomous underwater vehicles

Abstract

Advances in the miniaturization of microelectronics has greatly contributed to the proliferation of small, low cost autonomous underwater vehicles (AUVs). These affordable vehicles offer organizations a flexible platform that can be adapted to support a multitude of research goals. The small size and low entry cost come with a trade off of simple navigation systems, typically dead reckoning (DR) using a speed determined via propeller counts and heading from a low cost micro-electromechanical system (MEMS) inertial measurement unit (IMU), whose error grows unbounded without the availability of a ground referenced fix source and is compounded by the bias present in the speed measurement due to the change in hydrodynamics from the addition of sensors to the hull form. Additionally, some capabilities such as water current velocity measurement traditionally requires the addition of equipment that is not only expensive, but also whose size and power consumption can adversely affect operating characteristics and deployment times. This thesis expands on previous research using one-way travel time inverted USBL (OWTT-iUSBL) to calculate the local current velocity without the addition of a Doppler velocity log (DVL) or acoustic Doppler current profiler (ADCP). A novel extended Kalman filter (EKF) is proposed that, in addition to calculating the current velocity, estimates and corrects for the bias present in the speed measurement as determined by the main vehicle computer. Using data collected on the Charles River at the Massachusetts Institute of Technology (MIT) Sailing Pavilion, it is shown that current velocities can be reasonably calculated using OWTT-iUSBL data as compared to the values calculated using long baseline (LBL) data.