A method of removing outliers from position information based on SVR algorithm

Abstract

The present invention relates to a method of removing outliers from position information based on an SVR algorithm, and adopts the structural risk minimization principle, and in the process, first obtains historical position solution data of 1 time before AUV current time as training data set; then constructs the SVR regression model and determine the parameters of the SVR model to be constructed;then the relative position of AUV is solved by pure distance method and the condition parameters of AUV at the current time are substituted into SVR model to get the position prediction; and finally,whether the calculated position value is an outlier or not is determined by setting the trust degree; if the difference is greater than the set trust degree, the current calculated value is eliminated and the position value at the next time is returned; otherwise, the value is retained. The invention realizes the preprocessing of the position measurement data in the underwater acoustic positioning, reduces the error generated by the outliers and is beneficial to the subsequent filtering processing.

Description

technical field [0001] The invention relates to the field of elimination method of position information outliers, in particular to a method for elimination of position information outliers based on SVR algorithm. Background technique [0002] The pure distance method positioning in the USBL system measures the distance at multiple positions and combines the AUV heading information to calculate the position information. Through the AUV initial position p, the vertical distance d between the AUV and the underwater sound source emitter, the AUV's heading speed v and heading angle To calculate the position of AUV in real time. When positioning in USBL, due to the influence of underwater noise, reverberation, multi-path effect of sound propagation, Doppler effect, nonlinear distribution of sound velocity and other factors, positioning errors will inevitably occur, which will affect the accuracy of AUV navigation and positioning . Sometimes 1% to 5% or even as much as 10% to 2...

AI Technical Summary

Problems solved by technology

These methods can eliminate outliers to a certain extent, but if the parameters are not well selected, it is likely that the data processing results will not be convincing because of too severe distortion or cannot achieve the result of effectively eliminating outliers.

Images