Multidirectional bending sensor using capillary fibers and machine learning for real-time applications

In this article, the design and implementation of a bidirectional curvature sensor based on a fiber-optic interferometer are presented. The sensor structure was fabricated by fusing a capillary fiber fragment between single-mode fibers (SMFs), with the addition of a long end capillary to promote a l...

Descripción completa

Detalles Bibliográficos
Autores: Vanegas Tenezaca, Evelyn Dayanara, Galarza Galarza, Marko, Dauliat, Romain, Jamier, Raphael, Roy, Philippe, López-Amo Sáinz, Manuel
Tipo de recurso: artículo
Estado:Versión publicada
Fecha de publicación:2025
País:España
Institución:Universidad Pública de Navarra
Repositorio:Academica-e. Repositorio Institucional de la Universidad Pública de Navarra
OAI Identifier:oai:academica-e.unavarra.es:2454/54309
Acceso en línea:https://hdl.handle.net/2454/54309
Access Level:acceso abierto
Palabra clave:Bend
Capillary fiber
Curvature
Machine learning
Optical fiber sensor
Descripción
Sumario:In this article, the design and implementation of a bidirectional curvature sensor based on a fiber-optic interferometer are presented. The sensor structure was fabricated by fusing a capillary fiber fragment between single-mode fibers (SMFs), with the addition of a long end capillary to promote a long interferometric section, forming a Fabry-Perot (FP) cavity. Detailed analysis of the curvature data was carried out using machine learning techniques, allowing accurate classification of curvature in both directions of rotation. The experimental results showed excellent agreement (R2: 0.9998) with the predicted values. The sensor exhibits a maximum error of 1.9485°. This approach presents significant potential for applications requiring accurate real-time curvature measurements.