View in EDS

Application of stacked ensemble algorithm in pressure sensing of the identical weak FBG flexible skin.

Saved in:
Bibliographic Details
Title: Application of stacked ensemble algorithm in pressure sensing of the identical weak FBG flexible skin.
Authors: Fu, Guangwei1,2,3 (AUTHOR) earl@ysu.edu.cn, Liu, Ruida2 (AUTHOR), Zhang, Tiexin2 (AUTHOR), Zhang, Jiatong2 (AUTHOR), Jin, Wa2 (AUTHOR), Bi, Weihong1 (AUTHOR), Fu, Xinghu2 (AUTHOR)
Source: Measurement (02632241). Mar2026, Vol. 264, pN.PAG-N.PAG. 1p.
Subjects: Pressure sensors, Fiber Bragg gratings, Machine learning, Metaheuristic algorithms, Robotics, Ensemble learning
Abstract: • Predicting pressure signals of flexible skin using stacked ensemble algorithm. • Optimize the superposition algorithm hyperparameters and feature weights using WOA. • Demonstrated the potential of integrated algorithms in flexible skin sensing. Accurate positioning and measurement of force contact are crucial for flexible robot skin, enabling precise environmental interaction. However, current flexible pressure sensors have limitations in accurately predicting pressure. Most sensors use electrical sensors, and their dependence on electrical sensing mechanisms increases their sensitivity to electromagnetic interference, and the manufacturing process is complex. To address these issues, we use identical weak fiber Bragg grating to construct an optical flexible pressure sensor. Machine learning algorithms are used to achieve accurate prediction of force position and magnitude. This article studies a hybrid stacking algorithm that uses extremely randomized trees, Categorical Boosting, and multi-layer perceptron as the base models, and random forest as the meta model. In order to achieve more accurate pressure sensing, this paper introduces whale optimization algorithm to optimize the model hyperparameters and feature weights, improving the prediction accuracy of the algorithm. The optimized system achieved a mean absolute error of 1.12 mm and root mean squared error of 2.04 mm for low error sensing pressure position prediction. The mean absolute error for pressure magnitude prediction was 0.3N and root mean squared error was 0.45N. Our results demonstrate that optical frequency domain reflectometry-enabled identical weak fiber Bragg grating skins, combined with stacked ensemble learning, provide millimeter-level localization and accurate force estimation over a 40 × 40 mm area. The multiplexing capacity and flexible material make this approach attractive for robotic manipulation and wearable feedback. future work will investigate denser layouts, dynamic and multi-contact scenarios, aiming for sub-millimeter accuracy with real-time latency. [ABSTRACT FROM AUTHOR]
Copyright of Measurement (02632241) is the property of Elsevier B.V. and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database: Engineering Source
Description
Abstract:• Predicting pressure signals of flexible skin using stacked ensemble algorithm. • Optimize the superposition algorithm hyperparameters and feature weights using WOA. • Demonstrated the potential of integrated algorithms in flexible skin sensing. Accurate positioning and measurement of force contact are crucial for flexible robot skin, enabling precise environmental interaction. However, current flexible pressure sensors have limitations in accurately predicting pressure. Most sensors use electrical sensors, and their dependence on electrical sensing mechanisms increases their sensitivity to electromagnetic interference, and the manufacturing process is complex. To address these issues, we use identical weak fiber Bragg grating to construct an optical flexible pressure sensor. Machine learning algorithms are used to achieve accurate prediction of force position and magnitude. This article studies a hybrid stacking algorithm that uses extremely randomized trees, Categorical Boosting, and multi-layer perceptron as the base models, and random forest as the meta model. In order to achieve more accurate pressure sensing, this paper introduces whale optimization algorithm to optimize the model hyperparameters and feature weights, improving the prediction accuracy of the algorithm. The optimized system achieved a mean absolute error of 1.12 mm and root mean squared error of 2.04 mm for low error sensing pressure position prediction. The mean absolute error for pressure magnitude prediction was 0.3N and root mean squared error was 0.45N. Our results demonstrate that optical frequency domain reflectometry-enabled identical weak fiber Bragg grating skins, combined with stacked ensemble learning, provide millimeter-level localization and accurate force estimation over a 40 × 40 mm area. The multiplexing capacity and flexible material make this approach attractive for robotic manipulation and wearable feedback. future work will investigate denser layouts, dynamic and multi-contact scenarios, aiming for sub-millimeter accuracy with real-time latency. [ABSTRACT FROM AUTHOR]
ISSN:02632241
DOI:10.1016/j.measurement.2025.120215