P44 - Supervised Self-Calibration for Fault-Tolerant xMR-Based Angular Decoders under Dynamic Perturbations

Event
SMSI 2025
2025-05-06 - 2025-05-08
Nürnberg
Band
Poster
Chapter
Poster Session
Author(s)
E. Gerken, A. König - RPTU University Kaiserslautern Landau, Kaiserslautern (Germany)
Pages
319 - 320
DOI
10.5162/SMSI2025/P44
ISBN
978-3-910600-06-5
Price
free

Abstract

This paper introduces a self-calibration methodology inspired by biological systems, focusing on its ap-plication in xMR angular decoders. Designed to address the limitations of static calibration, the approach effectively corrects mechanical misalignments and operational deviations. By leveraging simulated and empirical data to train machine learning (ML) models such as support vector regression (SVR), convo-lutional neural networks (CNN), and resource-allocating networks (RAN) with radial basis function (RBF) components, it enables real-time error compensation. Experimental results demonstrate a reduction in mean absolute error (MAE) from 2.06° to 0.08° and confirm a significant improvement in recovery effi-ciency, robustness, and reliability.

Download