Hybrid Neural Network Models of Six-Axis Force Sensor

Abstract

A method of analyzing the Six-axis force measuring system by hybrid modeling is introduced in this paper. The mapping function of signal voltage output, which is input vectors of the Neural Network (NN) model, and measuring force signal, which is output vectors of the NN model, is represented as two parts. The determined linear part obtains the main principle and the the information of transfer matrix. The undetermined nonlinear part are estimated by neural network. The problems about nonlinear error and coupling are solved. The accuracy and feasibility of the method are displayed by the result of experiment data simulation.