Advancing Deformation-Sensitive Tactile Distribution Sensing through Electrical Impedance Tomography on Conductive Surfaces

Abstract

Electrical Impedance Tomography (EIT) is a non-invasive imaging technique that reconstructs the internal resistivity distribution of objects by analyzing electrical measurements obtained from boundary electrodes. This thesis advances the application of EIT by developing a measurement setup tailored for object localization and deformation-sensitive tactile sensing on conductive surfaces. The study investigates various excitation and measurement patterns, leveraging reconstruction algorithms from an open-source Python-based EIT library to enhance localization accuracy. A key challenge in EIT lies in its ill-posed inverse problem, where minor errors in boundary measurements can lead to significant distortions in reconstructed images. To address this, the research explores different electrode configurations and current injection patterns, optimizing them for spatial resolution, depth penetration, and robustness against noise. Three primary excitation patterns—cross, adjacent, and opposite—are systematically tested on homogeneous water phantoms and deformable conductive surface phantoms to evaluate their effectiveness. Experimental results demonstrate that pattern selection significantly impacts the accuracy of reconstructed images, with the cross and opposite patterns exhibiting superior depth sensitivity and reduced boundary artifacts compared to the adjacent pattern. Additionally, the study introduces a deformable conductive surface phantom, demonstrating the feasibility of EIT for tactile sensing applications. The findings suggest that EIT can effectively map impedance variations on flexible conductive surfaces, highlighting its potential for use in surgical simulators. The research further contributes by refining standard EIT algorithms to improve computational efficiency and image reconstruction fidelity. By optimizing measurement protocols and refining algorithmic approaches, this thesis enhances the practical applicability of EIT for real-world scenarios. The results provide valuable insights for future developments in medical imaging, industrial diagnostics, and sensor technology, paving the way for more accurate and reliable EIT-based sensing systems.

Date of Award	2025
Original language	English
Supervisor	Andreas Schrempf (Supervisor)