Publications & Projects
Humanoid Robot Whole-body Geometric Calibration with Embedded Sensors and a Single Plane
Abstract
This paper presents a novel method for whole-body geometric calibration of humanoid robots using a single plane, embedded sensors, and an admittance controller, introducing the IROC algorithm for optimal posture selection that achieved 2.3x accuracy improvement on the TALOS robot.
Improving Operational Accuracy of a Mobile Manipulator by Modeling Geometric and Non-Geometric Parameters
2024 IEEE-RAS 23rd International Conference on Humanoid Robots (Humanoids) - Nancy, France
Abstract
This paper improves mobile manipulator accuracy by modeling geometric and non-geometric parameters including base suspension and joint backlash, achieving 60% better end-effector pose estimation on the TIAGo robot.
FIGAROH: a Python toolbox for dynamic identification and geometric calibration of robots and humans
2023 IEEE-RAS 22nd International Conference on Humanoid Robots (Humanoids) - Austin, Texas
Abstract
This paper introduces FIGAROH, a Python toolbox for geometric calibration and dynamic identification of robots and humans that automates model building, data processing, and optimal trajectory generation.
Non-cylindrical wrapping rods and compliant adapters for multi-mode twisting string actuation
International Journal of Mechanisms and Robotic Systems 5 (1-2), 93-110
Abstract
This paper improves twisting string actuation efficiency through non-cylindrical wrapping rods and compliant adapters, demonstrating enhanced mechanical advantage and speed for robotic applications.
Recent Blog Posts
SO-ARM100 Full-Stack: From 3D Printing to Learning-Enabled Manipulation
May 14, 2025
Preparing for the LeRobot Global Hackathon 2025 meant building a complete SO-ARM100 stack—from 3D-printed parts all the way to learning-enabled manipulation demos. This guide distills what worked for me into a reusable playbook. It leans on the internal project notes in the fullstack-manip repository and should help you bring up your own SO-ARM, extend it to other manipulators, and plan a month-long sprint from research to evaluation.