Guilherme Christmann
  • Posts
  • Pubs
  • Pics

Publications

Benchmarking Smoothness and Reducing High-Frequency Oscillations in Continuous Control Policies teaser.
IROS 2024 Benchmarking Smoothness and Reducing High-Frequency Oscillations in Continuous Control Policies High-frequency oscillations are common in RL, we worked on how to improve smoothness without sacrifing performance.
Project Page
ArXiv
ArXiv
Expert Composer Policy - Scalable Skill Repertoire for Quadruped Robots teaser.
ICRA 2024 Expert Composer Policy - Scalable Skill Repertoire for Quadruped Robots Expert Composer Policy lets quadruped agents seamlessly chain motion skills with near-perfect transitions, enabling scalable and modular repertoire expansion.
Project Page
ArXiv
ArXiv
Expanding Versatility of Agile Locomotion through Policy Transitions Using Latent State Representation teaser.
ICRA 2023 Expanding Versatility of Agile Locomotion through Policy Transitions Using Latent State Representation A lightweight transition net lets quadruped robots switch between gaits in the real world with high success, enabling scalable skill composition without retraining.
Project Page
ArXiv
ArXiv