Training-free Controllable Human Motion Generation under Heterogeneous Constraints

2026-07-02

AI Summary

This work addresses a key limitation of existing training-free controllable motion generation methods, which are restricted to differentiable objective-type constraints and struggle with non-differentiable, sparse, or black-box criterion-type constraints. To overcome this, the authors propose the Motion-Inference-as-Control (MIaC) framework, which formulates diffusion-based motion generation as a stochastic control problem. MIaC is the first approach to unify both objective-type and criterion-type heterogeneous constraints without requiring retraining. It achieves dynamic conflict resolution among multiple constraints through a stepwise control policy, gradient-free constraint guidance, and an adaptive coordination mechanism. Experiments demonstrate that the method significantly improves motion quality and constraint satisfaction across diverse and complex constraint scenarios.

Abstract

Training-free controllable motion generation has attracted growing interest for enabling flexible constraint enforcement without constraint-specific training. However, existing training-free methods require constraints to be continuous objective-based with differentiable losses, while many real-world requirements are criterion-based and provide only discontinuous, sparse, or even black-box feedback. In this paper, we propose Motion-Inference-as-Control (MIC), the first training-free motion generation framework that handles both continuous objective-based and criterion-based motion constraints under a shared mechanism. The key idea is to cast diffusion-based motion generation as a stochastic control problem. This perspective not only provides principled and practically effective step-wise control laws that support criterion-based constraints without requiring differentiability and naturally accommodate objective-based constraints as a special case, but also motivates a control-oriented constraint coordination mechanism that adaptively balances and reconciles motion constraints during generation. Experiments across diverse constraint settings demonstrate the effectiveness of our framework.

Problem

• controllable motion generation
• training-free
• heterogeneous constraints
• criterion-based constraints
• diffusion models

Innovation

• training-free
• diffusion model
• stochastic control
• motion generation
• constraint coordination