FIRST STEPS TOWARD LEVERAGING ARTIFICIAL INTELLIGENCE FOR PRECISE CHARACTERISATION OF FORCE TRANSDUCERS

D. Mirian, R. Kumme, R. Tutsch

Abstract

This work is dedicated to the demonstration of a dynamic force measurement system for precise characterisation of the force transducers. The rocking motion of the system as a main dominant source of uncertainty in the acceleration is investigated. We propose a novel method based on the application of an artificial neural network for evaluation of the data as an alternative to traditional approaches to get low-uncertainty calibration measurements. In the end, two special architectures of the artificial neural network, namely Long Short-Term Memory LSTM and Gated Recurrent Network GRU are introduced, and their appropriateness for our use case is discussed.

Keywords

dynamic force calibration; rocking motion; measurement uncertainty; machine learning; deep learning; recurrent neural networks

Download

IMEKO-TC3-2022-075.pdf

DOI

10.21014/tc3-2022.075

IMEKO TC

TC3 - Measurement of Force, Mass, Torque, and Gravity

Back to the proceedings