NaP-Control: Navigating Diffusion Prior for Versatile and Fast Character Control

2026-04-15

arXiv.org

AI Summary

This work addresses the limitations of existing physics-driven character animation methods based on diffusion models, which rely on gradient-based guidance during inference, resulting in slow execution and insufficient robustness. To overcome these issues, the authors propose a novel mechanism that eliminates the need for iterative test-time guidance by directly modulating the latent noise of a task-agnostic diffusion policy prior through reinforcement learning. During training, the approach integrates online environment interaction to optimize task-specific rewards, enabling the generation of high-fidelity motions tailored to particular tasks. The proposed method significantly improves inference speed and task success rates while preserving naturalness and fluidity of motion in complex scenarios, demonstrating superior performance across a variety of tasks.

Abstract

Achieving precise, versatile whole-body character control in physics-based animation remains challenging. Recent diffusion-based policies generate rich and expressive motions but typically rely on gradient-based test-time guidance to satisfy task objectives, which is slow and can reduce robustness. We introduce NaP-Control (Navigating Diffusion Prior for Versatile and Fast Character Control), abbreviated as NaP. Our method uses reinforcement learning to manipulate the latent noise of a task-agnostic diffusion policy prior, steering it toward task-specific behaviors for fast, robust control with high motion fidelity. In contrast to methods that rely solely on offline training, NaP interacts with the environment during training to correct motions and optimize task rewards, improving success rates and enabling adaptation to challenging scenarios. By directly predicting task-optimized diffusion noise, NaP eliminates iterative guidance during denoising and enables efficient inference. Experiments show that NaP attains higher success rates and faster inference while preserving natural motion across diverse tasks.

Problem

• character control
• diffusion models
• physics-based animation
• test-time guidance
• motion fidelity

Innovation

• diffusion policy
• reinforcement learning
• character control
• latent noise manipulation
• physics-based animation