Three-axis flexible tube sensor with LSTM-based force prediction for alignment of electric vehicle charging ports

This paper introduces a novel three-axis flexible tube sensor designed for force measurement in electric vehicle (EV) charging port alignment, utilizing long short-term memory (LSTM) networks. The research aims to develop and validate a flexible and accurate sensor system capable of predicting multi-axis forces during alignment. The sensor integrates a magnetic sensor at the center of a flexible tube to capture three-dimensional (3-D) magnetic field variations corresponding to force changes. Fabricated using thermoplastic polyurethane (TPU) via 3-D printing technology, the sensor leverages machine learning to predict force values along the , , and axes ( , , ). Finite element method (FEM) analysis was conducted to assess the deflection characteristics of the flexible tube under various force conditions. Experimental results demonstrate that integrating LSTM significantly enhances the accuracy of force prediction, achieving an R² score exceeding 97 % for all axes, with mean squared error (MSE) values of 0.2819 for the -axis, 0.3567 for the -axis, and 2.8086 for the -axis. The sensor is capable of measuring forces up to 30 N without exceeding its elastic limits. These findings highlight the sensor’s potential for improving alignment accuracy and reliability in automated EV charging systems.