LEGO: Latent‑space Exploration for Geometry‑aware Optimization of Humanoid Kinematic Design

1Korea University, 2Contoro Robotics, 3MIT

Abstract

Designing robot morphologies and kinematics has traditionally relied on human intuition, with little systematic foundation. Motion–design co-optimization offers a promising path toward automation, but two major challenges remain: (i) the vast, unstructured design space and (ii) the difficulty of constructing task-specific loss functions.

We propose a new paradigm that minimizes human involvement by:

  • Learning the design search space from existing mechanical designs, rather than hand-crafting it.
  • Defining the loss directly from human motion data via motion retargeting and Procrustes analysis.

Using screw-theory-based joint axis representation and isometric manifold learning, we construct a compact, geometry-preserving latent space of robot designs in which optimization is tractable. We then solve design optimization in this latent space using gradient-free optimization. Our approach establishes a principled framework for data-driven robot design and demonstrates that leveraging existing designs and human motion can effectively guide the automated discovery of novel robot design.

Screw Theory Based Representation

Our framework models each joint as a 6D vector S=(ω,q) , where ω is the unit axis and q is the global position , enabling a unified and globally consistent description of diverse humanoid morphologies that preserves explicit positional information for motion retargeting.

g1

hrp4

igus_op

Latent Space Optimization

Explore and manipulate the generated robot models by our method.

BibTeX

@article{yoon2026LEGO, 
  title={LEGO: Latent‑space Exploration for Geometry‑aware Optimization of Humanoid Kinematic Design}, 
  author={JiHwan Yoon, Taemoon Jeong, Jeongeun Park, Chanwoo Kim, 
    Jaewoon Kwon, Yonghyeon Lee, Kyungjae Lee, Sungjoon Choi}, 
  journal={arxiv(will be uploaded)}, 
  year={2026} 
}