Improving Strain-Sensing Electronic Skins Based on Electrical Impedance Tomography
Date
Authors
Supervisor
Item type
Degree name
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Background: Electrical impedance tomography (EIT) possesses the capability of converting almost any electrically conductive material into an electronic skin (E-skin) that detects and transduces spatially distributed pressures and strains. In recent years, EIT-based E-skins have gained considerable research interest due to their potential applications in wearable devices, human-machine interfaces, soft robotics, and prosthetics. However, their practicality is limited due to a severe lack of accuracy.
Aim: This research aims to discover new strategies for integrating EIT with flexible strain-transducing materials to realize an EIT-based E-skin that offers improved performance in sensing two-dimensional strain distribution. Towards this goal, the critical limitations possessed by the existing approach for realizing EIT-based E-skins are mitigated by adopting a novel structural model, constraining the solution space, and utilizing a material that provides consistent strain-signal transduction.
Method: First, the limitations of various strain-sensing materials and conventional reconstruction algorithms were analysed via a literature review and a simulation study. Second, the reconstruction algorithms of EIT were adapted by replacing the conventional structural model with a new model inspired by the nodal admittance matrix (NAM) and constraining the solution space according to material properties. The effectiveness of the proposed reconstruction method was verified by an additional simulation study. Third, comprehensive material testing was performed to examine the strain-sensing characteristics of two candidate materials and allow the selection of a more suitable material. Finally, the selected material was used to fabricate an E-skin sensor, whose performance was validated experimentally.
Results: The experimental results showed that the sensor successfully achieved detecting, localizing, and transducing spatially distributed strains. The average percentage of mismatch between the applied strains and the reconstructed strains was less than 20%.