Underwater robot kinetic model parameter identification method based on Huber M estimation

Abstract

The invention discloses an underwater robot kinetic model parameter identification method based on Huber M estimation. The underwater robot kinetic model parameter identification method comprises thefollowing steps: 1) establishing an underwater robot kinetic model; 2) determining kinetic parameters required to be identified according to the established model; 3) adopting a recursive least squaremethod based on a Huber loss function to identify a parameter estimation value of the underwater robot kinetic model; 4) updating the current state value of the underwater robot according to the parameter estimation value of the kinetic model of the underwater robot obtained by identification; and (5) repeating the steps (3) to (4) to obtain a parameter estimation value at each sampling moment, and taking an average value as an identification result. According to the underwater robot kinetic model parameter identification method disclosed by the invention, the to-be-identified parameter can still be stably identified in an outlier noise environment, so that the identification precision is improved, and the robustness is improved.

Description

technical field [0001] The invention relates to a method for identifying model parameters, in particular to a method for identifying parameters of a dynamic model of an underwater robot based on Huber M estimation. Background technique [0002] Underwater interference is very complex. Simple Gaussian noise cannot fully describe the complex noise generated by water flow. The mean value of the noise cannot be guaranteed to be zero, and the standard deviation cannot always be constant. Many studies have shown that under the interference of simple Gaussian noise, the effect of the least squares method will be more obvious, but if a certain proportion of "outliers" (noise outliers with a large standard deviation) appear in the interference noise, the ordinary least square method is still used at this time. If the square method or the recursive least squares method is used to identify the model parameters, the estimation results will be inaccurate or even divergent, which will dir...

AI Technical Summary

Problems solved by technology

[0002] Underwater interference is very complex. Simple Gaussian noise cannot fully describe the complex noise generated by water flow. The mean value of the noise cannot be guaranteed to be zero, and the standard deviation cannot be constant.

Many studies have shown that under the interference of simple Gaussian noise, the effect of the least squares method will be more obvious, but if a certain proportion of "outliers" (noise outliers with a large standard deviation) appear in the interference noise, the ordinary least square method is still used at this time. The identification of model parameters by the square method or the recursive least squares method will lead to inaccurate or even divergent estimation results, which will directly lead to the deterioration of the robustness of model identification.

Images