Improved Transfer Learning for Detecting Upper-Limb Movement Intention Using Mechanical Sensors in an Exoskeletal Rehabilitation System.

The objective of this study was to propose a novel strategy for detecting upper-limb motion intentions from mechanical sensor signals using deep and heterogeneous transfer learning techniques. Three sensor types, surface electromyography (sEMG), force-sensitive resistors (FSRs), and inertial measurement units (IMUs), were combined to capture biometric signals during arm-up, hold, and arm-down movements. To distinguish motion intentions, deep learning models were constructed using the CIFAR-ResNet18 and CIFAR-MobileNetV2 architectures.

Experimental design of 4-point supported belt robot for sanding large convex surfaces.

In general, sanding robots that move as if drawing a line along a surface are mainly used when sanding objects with a large area; however, they require a long working time, and it is difficult to secure a uniform sanded area. This study focuses on large-area sanding robots, such as those for ships, storage tanks, and tank lorries, and proposes an adaptive belt tension robot equipped with a 4-point supported belt mechanism capable of sanding variable curved surfaces. In addition, a sanding normal force prediction formula is proposed to describe the sanding performance of the contact surface.