Register or Log In

International Society for Industrial Process Tomography

12th World Congress on Industrial Process Tomography

On-edge tactile perception with flexible electrical impedance tomography

Xiaopeng Wu1, Huazhi Dong1, Chengjia Wang2, Yunjie Yang1

1 SMART Group, School of Engineering, The University of Edinburgh, Edinburgh, United Kingdom
2 School of Mathematical and Computer Sciences, Heriot-Watt University, Edinburgh, United Kingdom

*Email: y.yang@ed.ac.uk


ABSTRACT

Closed-loop control is a critical component in the practical application of robots. To enable effective feedback and interaction with the environment, real-time and low-latency robotic perception is necessary. Traditional cloud-based approaches often suffer from delays caused by the need to transmit large volumes of data between local systems and remote servers. To overcome this limitation, we propose an edge computing-enabled tactile sensing framework, where data acquisition software and machine learning models are deployed directly on edge devices, thereby eliminating transmission delays and enhancing the system's responsiveness. In this work, we developed an edge computing system for Flexible Electrical Impedance Tomography (FEIT) on NVIDIA’s Jetson platform. It integrates several functionalities, including sensory signal acquisition, local model inference and real-time visualization of the tactile reconstruction results. Moreover, the system architecture emphasizes modularity and adaptability, allowing for fine-tuning and integration with diverse EIT reconstruction models, which broadens its applicability in robotic sensing and control. Tests have demonstrated that our EIT-based edge perception system performs


Keywords: Electrical Impedance Tomography, Tactile perception, Edge computing

Industrial Application: General

Sign-in to access the full text