Data-driven depth-averaged current prediction methods for underwater gliders with sailing parameters

The study of depth-averaged currents is great significance for the application underwater gliders. In order to solve problem low prediction accuracy time series-based current method, factors affecting are analyzed and a data-driven method an glider with sailing parameters proposed in this paper. First, glider’s historical profile period navigation taken as inputs construct multi-input double-output characteristics. Then, based on two sets real sea trial data groups generic set evaluation criteria, five different methods used predict currents. Experimental results show that result driven by better than series, future improved.