This thesis presents a low-cost proprioceptive and control framework for a soft robotic scraper by integrating optical and magnetic sensing, Kalman-based estimation, and deep-learning reconstruction. A redesigned fiber-based sensor enables accurate curvature estimation, while a hierarchical control scheme achieves stable shape servoing in realistic contact tasks. The proposed multimodal pipeline reliably reconstructs the scraper’s deformation using only embedded sensors, demonstrating an accessible and effective approach to soft-robot shape sensing and control.
Questa tesi presenta un sistema di propriocezione e controllo a basso costo per una spatola robotica soffice, integrando sensing ottico e magnetico, stima basata su Kalman e ricostruzione tramite deep learning. Un sensore a fibre ottiche ridisegnato consente un’accurata stima della curvatura, mentre un’architettura di controllo gerarchica garantisce un servoing stabile in condizioni di contatto reali. Il pipeline multimodale proposto ricostruisce in modo affidabile la deformazione della spatola utilizzando solo sensori integrati, dimostrando un approccio accessibile ed efficace al sensing e al controllo della forma nei robot soffici.
Design and integration of an optical fiber based scraper for multimodal shape sensing in soft robotics
Dinardo, Francesco
2024/2025
Abstract
This thesis presents a low-cost proprioceptive and control framework for a soft robotic scraper by integrating optical and magnetic sensing, Kalman-based estimation, and deep-learning reconstruction. A redesigned fiber-based sensor enables accurate curvature estimation, while a hierarchical control scheme achieves stable shape servoing in realistic contact tasks. The proposed multimodal pipeline reliably reconstructs the scraper’s deformation using only embedded sensors, demonstrating an accessible and effective approach to soft-robot shape sensing and control.| File | Dimensione | Formato | |
|---|---|---|---|
|
2025_12_Dinardo.pdf
non accessibile
Descrizione: Testo Tesi corretto
Dimensione
1.66 MB
Formato
Adobe PDF
|
1.66 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/247640