MotionPersona: Characteristics-aware Locomotion Control

2025-05-30

AI Summary

Existing character controllers struggle to model real-world human diversity, producing homogeneous motions that lack responsiveness to anatomical features, psychological states, and demographic attributes. To address this, we propose the first fine-grained trait-driven real-time gait control system. Our method introduces: (1) a multidimensional conditional framework jointly conditioning on SMPLX pose parameters, textual prompts, and user-defined control signals; (2) a few-shot character profiling technique based on short motion clips, overcoming representational limitations of text-only prompting; and (3) a block-wise autoregressive motion diffusion model, accompanied by a high-quality, publicly released dataset covering diverse traits and gait styles. Experiments demonstrate significant improvements over state-of-the-art methods in trait fidelity, motion quality, and stylistic diversity, while enabling real-time interactive generation. Code, data, and interactive demos are open-sourced.

Technology Category

• Humans and AI: Game Design — Virtual Humans, NPCs and Autonomous Characters
• Natural Language Processing: Generation
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work

Abstract

We present MotionPersona, a novel real-time character controller that allows users to characterize a character by specifying attributes such as physical traits, mental states, and demographics, and projects these properties into the generated motions for animating the character. In contrast to existing deep learning-based controllers, which typically produce homogeneous animations tailored to a single, predefined character, MotionPersona accounts for the impact of various traits on human motion as observed in the real world. To achieve this, we develop a block autoregressive motion diffusion model conditioned on SMPLX parameters, textual prompts, and user-defined locomotion control signals. We also curate a comprehensive dataset featuring a wide range of locomotion types and actor traits to enable the training of this characteristic-aware controller. Unlike prior work, MotionPersona is the first method capable of generating motion that faithfully reflects user-specified characteristics (e.g., an elderly person's shuffling gait) while responding in real time to dynamic control inputs. Additionally, we introduce a few-shot characterization technique as a complementary conditioning mechanism, enabling customization via short motion clips when language prompts fall short. Through extensive experiments, we demonstrate that MotionPersona outperforms existing methods in characteristics-aware locomotion control, achieving superior motion quality and diversity. Results, code, and demo can be found at: https://motionpersona25.github.io/.

Problem

• Generates diverse motions based on user-specified character traits
• Real-time control for animations reflecting real-world human motion variations
• Few-shot customization when textual prompts are insufficient

Innovation

• Block autoregressive motion diffusion model
• SMPLX parameters and textual prompts conditioning
• Few-shot characterization technique customization