RoboStriker: Hierarchical Decision-Making for Autonomous Humanoid Boxing

2026-01-30

AI Summary

This work aims to endow humanoid robots with human-level competitive intelligence and physical agility in highly dynamic, high-contact scenarios such as boxing. To this end, a three-tier hierarchical framework is proposed that decouples high-level strategic decision-making from low-level physical control, leveraging human motion data to construct a skill repertoire. Key innovations include the formulation of a topology-constrained latent skill manifold to restrict exploration to physically feasible actions, and the introduction of a novel Latent-Space Neural Fictitious Self-Play (LS-NFSP) mechanism that effectively stabilizes multi-agent adversarial training. The approach integrates single-agent motion imitation, Gaussian distribution-based hyperspherical projection regularization, and multi-agent reinforcement learning, achieving superior competitive performance in simulation and successfully transferring trained policies to real-world humanoid robots.

Technology Category

• Intelligent Robots: Manipulation
• Multiagent Systems: Adversarial Agents
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Economics, Online Markets and Human Computation: Uses of LLMs and GenAI for marketplace design, bidding, and strategic interactions

Abstract

Achieving human-level competitive intelligence and physical agility in humanoid robots remains a major challenge, particularly in contact-rich and highly dynamic tasks such as boxing. While Multi-Agent Reinforcement Learning (MARL) offers a principled framework for strategic interaction, its direct application to humanoid control is hindered by high-dimensional contact dynamics and the absence of strong physical motion priors. We propose RoboStriker, a hierarchical three-stage framework that enables fully autonomous humanoid boxing by decoupling high-level strategic reasoning from low-level physical execution. The framework first learns a comprehensive repertoire of boxing skills by training a single-agent motion tracker on human motion capture data. These skills are subsequently distilled into a structured latent manifold, regularized by projecting the Gaussian-parameterized distribution onto a unit hypersphere. This topological constraint effectively confines exploration to the subspace of physically plausible motions. In the final stage, we introduce Latent-Space Neural Fictitious Self-Play (LS-NFSP), where competing agents learn competitive tactics by interacting within the latent action space rather than the raw motor space, significantly stabilizing multi-agent training. Experimental results demonstrate that RoboStriker achieves superior competitive performance in simulation and exhibits sim-to-real transfer. Our website is available at RoboStriker.

Problem

• humanoid robotics
• competitive intelligence
• physical agility
• contact-rich dynamics
• autonomous boxing

Innovation

• Hierarchical Reinforcement Learning
• Latent Action Space
• Neural Fictitious Self-Play
• Motion Manifold
• Sim-to-Real Transfer