Yosuke Komata

About

Yosuke Komata is a pioneering researcher in the field of biped locomotion robotics, with a primary focus on developing adaptive control systems that enable humanoid robots to achieve stable, natural walking motions. His major contributions center on the integration of evolutionary computation and neural network architectures to address the fundamental challenge of dynamic balance in bipedal movement. Komata’s most influential work, "Stabilization control of biped locomotion robot based learning with GAs having self-adaptive mutation and recurrent neural networks" (2002, 59 citations), introduced a novel method that combines self-adaptive genetic algorithms with recurrent neural networks to generate stable walking patterns in response to varying environments. This approach leverages force sensors at the robot’s soles to continuously adjust the zero moment point (ZMP)—a critical stability metric—in real time. In his related study "Recurrent neural network with self-adaptive GAs for biped locomotion robot" (2002, 17 citations), he further refined this framework, demonstrating how adaptive mutation rates in genetic algorithms can optimize neural network learning for more robust locomotion control. Komata’s work has been instrumental in advancing the field of humanoid robotics, providing foundational techniques that enable robots to navigate complex terrains with human-like efficiency.

Research Focus

Key Achievements

Top Papers

1. 1
   Stabilization control of biped locomotion robot based learning with GAs having self-adaptive mutation and recurrent neural networks

   59 citations · 2002
2. 2
   Recurrent neural network with self-adaptive GAs for biped locomotion robot

   17 citations · 2002

Key Collaborators

Contact & Links