Hiroshi Atsuta

My research focuses on the intersection of humanoid robotics and reservoir computing, driven by the core philosophy of embodied intelligence. I am interested in how physical bodies and computational frameworks can be co-optimized to produce natural, adaptive movement in complex machines.

Currently, my work centers on three primary pillars:

Humanoid Control & Locomotion: Developing data-driven inverse dynamics controllers for pneumatic humanoid robots and exploring omnidirectional control for bipedal locomotion.
Reservoir Computing (RC): Building efficient software frameworks to apply RC to robotics. I am the lead developer of rcl, a library designed to bring high-performance reservoir computing to real-time robotic control.
Simulation & Tooling: Creating specialized environments like skelarm, a dynamics simulator for planar robot arms, and frameworks for 7-DOF manipulator teaching and playback.

By bridging the gap between biological flexibility and artificial machine systems, I aim to create robots that do not just follow code, but interact fluidly with the physical world.