Flexible touch and gesture recognition system for curved surfaces with machine learning for assistive applications
Touch is a fundamental mode of human-machine interaction and ability to monitor tactile pressure, recognize gestures and location of touch are crucial for touch-based technologies. However, achieving reliable touch sensing on curved surfaces remains challenging as flexing often disrupts the stabilit...
Saved in:
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Elsevier
2025-06-01
|
Series: | Sensors and Actuators Reports |
Subjects: | |
Online Access: | http://www.sciencedirect.com/science/article/pii/S2666053925000049 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
_version_ | 1832591787880873984 |
---|---|
author | Gitansh Verma Shrutidhara Sarma Eugen Koch Andreas Dietzel |
author_facet | Gitansh Verma Shrutidhara Sarma Eugen Koch Andreas Dietzel |
author_sort | Gitansh Verma |
collection | DOAJ |
description | Touch is a fundamental mode of human-machine interaction and ability to monitor tactile pressure, recognize gestures and location of touch are crucial for touch-based technologies. However, achieving reliable touch sensing on curved surfaces remains challenging as flexing often disrupts the stability of sensor outputs and diminishes sensitivity, especially in dynamic environments. This study presents the development of a flexible multi-element touch sensing patch that can monitor its bending state as well as detect pressure with a sensitivity of 0.827 kPa−1. The patch is fabricated using resistive strain sensors, screen printed onto a PET sheet with a foam backing. Evaluation electronics were integrated to ensure stable, noise-free signal acquisition, and output was processed with machine learning (ML) algorithms to classify gestures such as single and double finger taps, swipes, and touch locations, with 93 % accuracy, on both flat and curved surfaces. Based on the identified gesture, the system enables users to type text or control external devices with minimal physical effort. Its scalable fabrication, high sensitivity, mechanical resilience and seamless ML integration establishes it as a powerful and efficient tool for assistive technologies, designed to support individuals with limited speech and mobility, such as those with quadriplegia or paralysis. |
format | Article |
id | doaj-art-ae75e0246d5d4a079e583169b654fe8e |
institution | Kabale University |
issn | 2666-0539 |
language | English |
publishDate | 2025-06-01 |
publisher | Elsevier |
record_format | Article |
series | Sensors and Actuators Reports |
spelling | doaj-art-ae75e0246d5d4a079e583169b654fe8e2025-01-22T05:44:05ZengElsevierSensors and Actuators Reports2666-05392025-06-019100284Flexible touch and gesture recognition system for curved surfaces with machine learning for assistive applicationsGitansh Verma0Shrutidhara Sarma1Eugen Koch2Andreas Dietzel3Department of Mechanical Engineering, IIT Jodhpur, Rajasthan 342030, IndiaDepartment of Mechanical Engineering, IIT Jodhpur, Rajasthan 342030, India; Institute of Microtechnology, TU Braunschweig, Alte Salzdahlumer Str. 203, Braunschweig 38124, Germany; Corresponding author: Institut für Mikrotechnik, Technische Universität Braunschweig, Alte Salzdahlumer Str. 203, 38124 Braunschweig, Germany.Institute of Microtechnology, TU Braunschweig, Alte Salzdahlumer Str. 203, Braunschweig 38124, GermanyInstitute of Microtechnology, TU Braunschweig, Alte Salzdahlumer Str. 203, Braunschweig 38124, GermanyTouch is a fundamental mode of human-machine interaction and ability to monitor tactile pressure, recognize gestures and location of touch are crucial for touch-based technologies. However, achieving reliable touch sensing on curved surfaces remains challenging as flexing often disrupts the stability of sensor outputs and diminishes sensitivity, especially in dynamic environments. This study presents the development of a flexible multi-element touch sensing patch that can monitor its bending state as well as detect pressure with a sensitivity of 0.827 kPa−1. The patch is fabricated using resistive strain sensors, screen printed onto a PET sheet with a foam backing. Evaluation electronics were integrated to ensure stable, noise-free signal acquisition, and output was processed with machine learning (ML) algorithms to classify gestures such as single and double finger taps, swipes, and touch locations, with 93 % accuracy, on both flat and curved surfaces. Based on the identified gesture, the system enables users to type text or control external devices with minimal physical effort. Its scalable fabrication, high sensitivity, mechanical resilience and seamless ML integration establishes it as a powerful and efficient tool for assistive technologies, designed to support individuals with limited speech and mobility, such as those with quadriplegia or paralysis.http://www.sciencedirect.com/science/article/pii/S2666053925000049Screen printingTouch and bending sensing patchMachine LearningEvaluation Electronics |
spellingShingle | Gitansh Verma Shrutidhara Sarma Eugen Koch Andreas Dietzel Flexible touch and gesture recognition system for curved surfaces with machine learning for assistive applications Sensors and Actuators Reports Screen printing Touch and bending sensing patch Machine Learning Evaluation Electronics |
title | Flexible touch and gesture recognition system for curved surfaces with machine learning for assistive applications |
title_full | Flexible touch and gesture recognition system for curved surfaces with machine learning for assistive applications |
title_fullStr | Flexible touch and gesture recognition system for curved surfaces with machine learning for assistive applications |
title_full_unstemmed | Flexible touch and gesture recognition system for curved surfaces with machine learning for assistive applications |
title_short | Flexible touch and gesture recognition system for curved surfaces with machine learning for assistive applications |
title_sort | flexible touch and gesture recognition system for curved surfaces with machine learning for assistive applications |
topic | Screen printing Touch and bending sensing patch Machine Learning Evaluation Electronics |
url | http://www.sciencedirect.com/science/article/pii/S2666053925000049 |
work_keys_str_mv | AT gitanshverma flexibletouchandgesturerecognitionsystemforcurvedsurfaceswithmachinelearningforassistiveapplications AT shrutidharasarma flexibletouchandgesturerecognitionsystemforcurvedsurfaceswithmachinelearningforassistiveapplications AT eugenkoch flexibletouchandgesturerecognitionsystemforcurvedsurfaceswithmachinelearningforassistiveapplications AT andreasdietzel flexibletouchandgesturerecognitionsystemforcurvedsurfaceswithmachinelearningforassistiveapplications |