Prediction method and system for deep average flow of underwater glider

Abstract

The invention provides a prediction method and system for deep average flow of an underwater glider. The method comprises the following steps: S10, acquiring an original deep average flow time sequence; S20, dividing the original deep average flow time sequence into a training data sequence and a to-be-predicted data sequence; S30, constructing an ARMA model and an LSSVM model, and training the ARMA model and the LSSVM model through the training data sequence to obtain an ARMA prediction model and an LSSVM prediction model; and S40, inputting a to-be-predicted data sequence into the ARMA prediction model and the LSSVM prediction model to obtain a prediction result. The invention has the beneficial effects of practicability and accuracy, and is suitable for the field of ocean observation.

Description

technical field [0001] The invention relates to the technical field of ocean observation, in particular to a method and system for predicting deep mean flow of an underwater glider. Background technique [0002] As a marine robot, the underwater glider is very sensitive to convection due to its buoyancy drive; during the voyage, due to the needs of observation tasks, it is required to perceive the ocean current in the local environment in advance, but the current ocean current forecast cannot provide it. It can effectively predict ocean currents mainly because: [0003] 1) Ocean current forecasting does not have large-scale and continuous in-situ observations, resulting in large uncertainties in initial values ​​and boundary conditions. The errors caused by these uncertainties in the forecasting process are gradually amplified, thus affecting the accuracy of ocean current forecasting. precision; [0004] 2) Limited by model prediction technology, storage technology and in-...

AI Technical Summary

Problems solved by technology

[0002] As a marine robot, the underwater glider is very sensitive to convection due to its buoyancy drive; during the voyage, due to the needs of observation tasks, it is required to perceive the ocean current in the local environment in advance, but the current ocean current forecast cannot provide it. It can effectively predict ocean currents mainly because:

[0003] 1) Ocean current forecasting does not have large-scale and continuous in-situ observations, resulting in large uncertainties in initial values ​​and boundary conditions. The errors caused by these uncertainties in the forecasting process are gradually amplified, thus affecting the accuracy of ocean current forecasting. precision;

[0004] 2) Limited by model prediction technology, storage technology and in-situ observation technology, the temporal and spatial resolution of ocean current forecasting is not very high, and it cannot fully meet the application requirements of underwater gliders for the time being.

[0006] Existing deep mean flow prediction methods lack practicability or have low accuracy. For example, the deep mean flow prediction method based on decomposition-combination decomposes the training data and the data to be predicted at one time in advance, which leads to its lack of practicability. If only the decomposition Training data, the prediction accuracy of existing methods is greatly reduced, which cannot meet the application requirements of underwater gliders

Images