Artificial Intelligence and Data Science
Online ISSN : 2435-9262
Potential for Real-World Applications of Imitation Learning Using a Multi-Fingered Robotic Hand and Tactile Sensing
Satoshi FUNABASHI
Author information
JOURNAL OPEN ACCESS

2026 Volume 7 Issue 3 Pages 1-13

Details
Abstract

Multi-fingered robotic hands, designed to achieve human-like manipulation with five fingers, are attract-ing growing interest for applications in labor-shortage mitigation and disaster response. We are developing Phase-Free Robots that can operate in both everyday and emergency situations. Inspired by the structure and function of the human hand and brain, recent advances in imitation learning using tactile sensors, tele-operation gloves, and VLA (Vision-Language-Action) models are being applied to multi-fingered hands. However, collecting high-quality human manipulation data remains challenging. We are pursuing practical multi-fingered robotic hands while exploring imitation learning for complex human-like manipulation.

Content from these authors
© 2026 Japan Society of Civil Engineers
Next article
feedback
Top