Keita Tokuda

Osaka Prefecture University, Kobe University, Wakayama University, Nagoya Institute of Technology

Papers

Total Citations

H-Index

About

Keita Tokuda’s research career is defined by a singular, ambitious focus: creating robots that can save lives in the aftermath of natural disasters. His foundational work on the “general rescue robot CUL” (2003, 11 citations) tackled the critical challenge of deploying machinery into environments where roads are destroyed and terrain is treacherous. To make such robots viable, Tokuda pioneered methods for environmental perception, most notably developing a novel foot pressure sensor for legged robots (2004, 8 citations). This sensor allows a robot to “feel” the ground beneath it, measuring reactive forces to estimate the stability of fragile rubble—a crucial capability for safe navigation in collapsed structures. Beyond mechanics, Tokuda advanced the software side of robotics. He created the FDNet dynamic programming environment (2005, 3 citations), a flexible, XML-based system that allows robot networks to be reconfigured on the fly to accept new functions. Demonstrating a holistic view of human-robot interaction, he also developed a Minimum Classification Error training algorithm (2006, 2 citations) enabling robots to learn and identify speakers during live interaction, a key step for effective communication in chaotic rescue scenes.