A Novel Physical HumanRobot Interface With Pressure Distribution Measurement Based on Electrical Impedance Tomography

Abstract

Measurements of pressure distribution on physical human-robot interfaces are crucial for ensuring the comfort and safety of human-exoskeleton interactions. To address this need, we propose a novel sensorized physical human-robot interface with pressure distribution measurement based on electrical impedance tomography (EIT). EIT sensors have the advantages of a simple structure and the ability to continuously measure pressure over a large area, making it a promising solution for wearable robots. A piezoresistive composite made of carbon black (CB) and flexible polymer was fabricated and embedded in a physical interface to compose the EIT sensor. To improve the spatial solution of the EIT inverse problem, a convolutional neural network (CNN)-enhanced Tikhonov regularization (referred to as CNN-TR) approach was adopted. The original compressed image is first reconstructed with Tikhonov regularization and then enhanced by the CNN model that is trained with simulation data. Then, a validation platform is built based on an MARK-10 force tester. The experimental results showed that the developed EIT sensor with CNN-TR reconstruction method achieves accurate localization and size estimation of the compressed area and is capable of distinguishing multiple compressed areas. In addition, the EIT sensor is embedded into a physical interface to measure the compressed areas between the physical interface and the human lower limb. The results validate that the proposed EIT pressure sensor is suitable for physical interfaces.