Section: New Results
Posture and motion capture by smart textile
The objective of the work is to design a jacket made of smart textile, without the use of built-in sensors, to determine the posture of the operator.
We propose an innovative solution based on the electrical properties of a stretchable conductive tissue which is used in the manufacture of a smart garment. We use the Electrical Impedance Tomography (EIT) to reconstruct the resistance change of the conductive tissue during tissue extension/deformation caused by human movement. The conductive tissue is placed at strategic points of the jacket (e.g., elbow, shoulder). The model that describes the correlation between the operator's posture/motion and tissue deformation is difficult to obtain analytically. Neural networks are being used to associate the different postures and movements measured by the reference device with the electric field measured in the smart textile. After the learning phase, the neural network is able to predict articular angle with an accuracy of + -5 degrees from tissue extension/deformation only.
Following the successful validation on the first prototype, a request of the patent was drafted and submitted on November 6, 2018 under the number FR1860192 (Smart textile adapted for motion and/ or deformation detection). At the same time, we submitted an experiment project to COERLE. The experiments are planned for next year. This study will allow us to acquire a big database for the learning of artificial neural networks in order to try to propose a unique and stable solution of human posture capture by the smart textile, whatever the anthropometric parameters.