Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

RoboForge: Physically Optimized Text-guided Whole-Body Locomotion for Humanoids

Xichen Yuan, Zhe Li, Bofan Lyu, Kuangji Zuo, Yanshuo Lu, Gen Li, Jianfei Yang

Abstract

While generative models have become effective at producing human-like motions from text, transferring these motions to humanoid robots for physical execution remains challenging. Existing pipelines are often limited by retargeting, where kinematic quality is undermined by physical infeasibility, contact-transition errors, and the high cost of real-world dynamical data. We present a unified latent-driven framework that bridges natural language and whole-body humanoid locomotion through a retarget-free, physics-optimized pipeline. Rather than treating generation and control as separate stages, our key insight is to couple them bidirectionally under physical constraints.We introduce a Physical Plausibility Optimization (PP-Opt) module as the coupling interface. In the forward direction, PP-Opt refines a teacher-student distillation policy with a plausibility-centric reward to suppress artifacts such as floating, skating, and penetration. In the backward direction, it converts reward-optimized simulation rollouts into high-quality explicit motion data, which is used to fine-tune the motion generator toward a more physically plausible latent distribution. This bidirectional design forms a self-improving cycle: the generator learns a physically grounded latent space, while the controller learns to execute latent-conditioned behaviors with dynamical integrity.Extensive experiments on the Unitree G1 humanoid show that our bidirectional optimization improves tracking accuracy and success rates. Across IsaacLab and MuJoCo, the implicit latent-driven pipeline consistently outperforms conventional explicit retargeting baselines in both precision and stability. By coupling diffusion-based motion generation with physical plausibility optimization, our framework provides a practical path toward deployable text-guided humanoid intelligence.

RoboForge: Physically Optimized Text-guided Whole-Body Locomotion for Humanoids

Abstract

While generative models have become effective at producing human-like motions from text, transferring these motions to humanoid robots for physical execution remains challenging. Existing pipelines are often limited by retargeting, where kinematic quality is undermined by physical infeasibility, contact-transition errors, and the high cost of real-world dynamical data. We present a unified latent-driven framework that bridges natural language and whole-body humanoid locomotion through a retarget-free, physics-optimized pipeline. Rather than treating generation and control as separate stages, our key insight is to couple them bidirectionally under physical constraints.We introduce a Physical Plausibility Optimization (PP-Opt) module as the coupling interface. In the forward direction, PP-Opt refines a teacher-student distillation policy with a plausibility-centric reward to suppress artifacts such as floating, skating, and penetration. In the backward direction, it converts reward-optimized simulation rollouts into high-quality explicit motion data, which is used to fine-tune the motion generator toward a more physically plausible latent distribution. This bidirectional design forms a self-improving cycle: the generator learns a physically grounded latent space, while the controller learns to execute latent-conditioned behaviors with dynamical integrity.Extensive experiments on the Unitree G1 humanoid show that our bidirectional optimization improves tracking accuracy and success rates. Across IsaacLab and MuJoCo, the implicit latent-driven pipeline consistently outperforms conventional explicit retargeting baselines in both precision and stability. By coupling diffusion-based motion generation with physical plausibility optimization, our framework provides a practical path toward deployable text-guided humanoid intelligence.
Paper Structure (22 sections, 7 equations, 3 figures, 4 tables)

This paper contains 22 sections, 7 equations, 3 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Related works
  3. Text-to-Motion Generation and Latent Diffusion
  4. Physics-Grounded Generation and Refinement
  5. Humanoid Motion Tracking and Retarget-free Inference Pipeline
  6. Positioning of Our Work
  7. METHODOLOGY
  8. Problem Formulation
  9. Motion Latent Diffusion-Based Human Motion Generation
  10. Physics-Plausibility Optimization Module
  11. Forward optimization of the tracker
  12. Motion quality control and Backward fine tune
  13. Experiments
  14. Experimental Settings
  15. Datasets
  16. ...and 7 more sections

Figures (3)

  • Figure 1: RoboForge. aims to perform physically plausible motions.(a) performed a series of three martial arts kicks. (b) turned around and threw the javelin with great force. (c) squats to avoid, then kicks, and finally punches.
  • Figure 2: The RoboForge end-to-end framework. The framework is composed of four key modules: (i) Motion Generation module, (ii) Physical Plausibility Optimization Module, (iii) Training Pipeline, and (iv) Sim-to-Real Deployment.
  • Figure 3: Text-guided examples in simulation. The humanoid performs diverse motions including boxing, jumping jacks, and martial arts kicking under different text prompts.