Host: The Japan Society of Mechanical Engineers
Name : [in Japanese]
Date : June 28, 2023 - July 01, 2023
Flexible actuators are popular in the consumer and medical fields because of their flexibility and compliance. However, they are typically difficult to model because of their viscoelasticity and nonlinearity. This paper presents an application of Transformer-based point cloud auto-encoder to catch the feature of fabric-type actuator using point cloud in high dimension, and evaluated by reconstruction task. The proposed method employs an asymmetric design and a shifting mask tokens operation to learn high-level latent features from unmasked point patches. The results show that the proposed approach achieved significant accuracy in reconstructing both real and simulated point cloud data. The proposed method has potential applications in wearable devices, soft grippers, and other soft robotic systems.
