Nonlinear observer for depth-aided INS: Experimental evaluation using an AUV

This paper presents the experimental results from testing a nonlinear observer on data obtained by an Autonomous Underwater Vehicle (AUV). The observer is modified to only use a pressure sensor instead of position and velocity measurements obtained from a Global Navigation Satellite System (GNSS) receiver. The observer utilizes the pressure sensor as a measurement of depth and the observer is therefore modified to only estimate attitude and vertical translational dynamics in the Earth-fixed North-East-Down (NED) reference frame. Through experimental testing the depth-aided observer shows excellent performance and comparable performance to the onboard Kalman filter. The nonlinear observer is global exponentially stable and the computational footprint is much smaller than a conventional Kalman filter.