Inversion of the sound speed profiles with an AUV carrying source using improved ensemble Kalman filter

This paper presents an improved ensemble Kalman filter with an autonomous underwater vehicle carrying source for sound speed profiles inversion. The Markov chain Monte Carlo method is introduced to improve the variability of ensemble members in the traditional ensemble Kalman filter and thus, improve the performance. The inverse problem is formulated as a state-space model with a state equation for the time-evolving sound speed profile and a measurement equation that incorporates acoustic measurements via a hydrophone array. Empirical orthogonal functions are introduced to reduce the unknown parameters in inversion algorithm. Simulation and experiment results show the proposed method correctly inverses the SSP and outperforms the traditional ensemble Kalman filter.