A Trajectory Prediction Method for Drifting Buoy Based on GA-DNN Model.

The floating movement of the buoy is a complex movement under the action of the Marine environment background field. In this paper, a deep neural network model (GA-DNN) optimized by a genetic algorithm is constructed based on the environmental factors affecting buoy drift, such as surface current and profile current, and the trajectory prediction of buoy drift is realized. Through a series of experiments, the GA-DNN model constructed in this paper has better prediction performance than the traditional deep neural network and linear regression model. The three evaluation indexes of RMSE, MAE, and R2 and distance and angle errors are improved, and the effective prediction of buoy trajectory is realized.