SceneBot: A unified humanoid tracking framework that combines free-space and contact-rich motion control

[Submitted on 25 Jun 2026]

8d ago· 2 min readenNews

technology science artificial intelligence robotics

Summary

SceneBot is a unified motion-tracking framework for humanoid robots that handles both free-space locomotion and contact-rich tasks (like carrying a box upstairs or traversing uneven terrain). Unlike current reinforcement-learning policies that struggle with physical ambiguities during object interaction, SceneBot conditions a single policy on both reference motions and per-link contact labels. To address the lack of annotated interaction data, the authors propose a hindsight scene reconstruction approach that infers scene-interaction graphs from retargeted human motion. Trained on 7.5 hours of reconstructed contact-rich data, SceneBot generalizes to unseen motions and environments, and is presented as the first general framework to unify free-space and contact-rich behaviors for humanoid control.

Source

Twitter / XSceneBot: A unified humanoid tracking framework that combines free-space and contact-rich motion controlarxiv.org

Key quotes

· 5 pulled

Current humanoid reinforcement-learning policies excel at free-space motions but struggle with contact-rich tasks, as pure kinematic tracking cannot resolve the physical ambiguities of interacting with objects and uneven terrain.

SceneBot conditions a single policy on both reference motions and per-link contact labels, explicitly defining expected environmental interactions.

To overcome the lack of annotated interaction data, we propose a hindsight scene reconstruction approach that infers scene-interaction graphs from retargeted human motion.

Our results demonstrate that SceneBot is the first general framework to seamlessly unify free-space and contact-rich behaviors executing complex, long-horizon tasks like carrying a box upstairs.

We establish contact conditioning as a powerful interface for humanoid control.

Snippet from the RSS feed

Current humanoid reinforcement-learning policies excel at free-space motions but struggle with contact-rich tasks, as pure kinematic tracking cannot resolve the physical ambiguities of interacting with objects and uneven terrain. To address this, we intro

You might also wanna read

ReMoT: A Reinforcement Learning Framework Using Motion Contrast Triplets to Improve VLM Spatio-Temporal Reasoning

ReMoT (Reinforcement Learning with Motion Contrast Triplets) is a unified training paradigm designed to address spatio-temporal consistency

arxiv.org·26d ago

Using Vision-Language Models to Segment Robot Demonstration Videos into Subtask Annotations

This article presents a benchmark and field report on using Vision-Language Models (VLMs) to segment robot demonstration videos and egocentr

macrodata.co·8d ago

DILLO: A Language-Based World Model for Proactive Agent Steering Without Visual Simulation

This paper introduces DILLO (DIstiLLed Language-ActiOn World Model), a proactive agent steering framework that replaces slow visual simulati

arxiv.org·15d ago

DynaFLIP: A Dynamics-Aware Multimodal Pre-Training Framework for Robot Manipulation Perception

DynaFLIP is a dynamics-aware multimodal pre-training framework for robot manipulation perception. It constructs image-language-3D flow tripl

arxiv.org·1mo ago

BilliardPhys-Bench: New Benchmark Reveals Physical Reasoning Gaps in Multimodal AI Models

This paper introduces BilliardPhys-Bench, a benchmark designed to evaluate multimodal large language models (MLLMs) on intuitive physical re