Koji Sokabe

Papers

Total Citations

H-Index

About

Koji Sokabe’s research lies at the intersection of human-robot collaboration, autonomous systems, and distributed sensing, with a strong emphasis on real-world applications such as agriculture and everyday human environments. His most influential work introduces a novel control architecture that leverages Gaussian Process (GP) regression to enable variable autonomy—allowing robotic systems to dynamically adjust their level of independence based on task complexity and human input. This framework has been applied to collaborative tasks like 3-D manual reaching operations, where humans and robots share control seamlessly. Sokabe also contributed to the Robot Town Project, developing distributed sensor networks—integrating cameras, laser range finders, and IC tags—for simultaneous multi-target tracking using advanced particle filters (SIR/MCMC). His work is foundational for creating safe, adaptive robots that operate alongside people in unstructured settings. With over a decade of research, Sokabe’s papers have garnered citations from the robotics and automation communities, reflecting his impact on semi-autonomous systems and human-robot interaction design. His achievements include pioneering variable autonomy methods that balance human oversight with robotic efficiency, a critical step toward practical, collaborative robotics.