DreamMimic: Learning Visuomotor Whole-Body Loco-Manipulation via World Model

• URL: http://arxiv.org/abs/2608.22278v1
• Date: Sun, 23 Aug 2026 08:08:39 GMT
• Title: DreamMimic: Learning Visuomotor Whole-Body Loco-Manipulation via World Model
• Abstract summary: Vision-based whole-body loco-manipulation on humanoid robots is challenging due to partial observability, contact-rich dynamics, and the difficulty of learning long-horizon behaviors from high-dimensional visual inputs.<n>We present a framework that distills privileged teacher policies into vision-based humanoid controllers via world-model-assisted distillation.
• Score: 5.082137967180834
• License: http://creativecommons.org/licenses/by-sa/4.0/
• Abstract: Vision-based whole-body loco-manipulation on humanoid robots is challenging due to partial observability, contact-rich dynamics, and the difficulty of learning long-horizon behaviors from high-dimensional visual inputs. We present \href{https://github.com/DreamMimic/DreamMimic}{DreamMimic}, a framework that distills privileged teacher policies into vision-based humanoid controllers via world-model-assisted distillation. Instead of using a Dreamer-style RSSM for planning, we repurpose it to learn predictive latent dynamics that serve as both a representation space and an action-conditioned multi-step supervision signal, while exposing compact predictive features to the student policy to reduce long-term drift. Beyond standard reconstruction objectives for proprioceptive and visual observations, we add auxiliary prediction heads for privileged state, contact, object state, and reward estimation. These heads provide additional supervision related to agent--object interaction and task progress, encouraging the latent representation to retain signals that are useful for contact-rich loco-manipulation. We further introduce Performance-Conditioned Guidance (PCG), a reward-driven adaptive distillation schedule that computes performance scores for both teacher and student to dynamically balance guidance and exploration. PCG prevents both premature teacher annealing and excessive teacher interference in challenging visual settings. Experiments on OMOMO and BEHAVE show improved tracking-based loco-manipulation performance over strong vision-based baselines, without exposing online privileged interaction states to the student at deployment. Qualitative simulations further examine morphology and simulator changes. These results suggest that world models can provide a useful mechanism for stabilizing visual policy distillation in contact-rich humanoid behaviors.

Related papers

• Towards Generalizable Visually Grounded Exploration of Household Devices [74.66117923766632]
  Key bottleneck of realizing general autonomous embodied agents lies in Generalizable Visually Grounded Exploration.<n>We introduce VGEBench, a benchmark designed to evaluate the generalizable visually grounded exploration capabilities of Vision-Language Models.<n>Unlike static datasets, we construct a Logic-Driven State Machine framework. This framework simulates multi-turn interaction loops, compelling agents to achieve goals by active visual perception and feedback-driven correction.
  arXiv  Detail & Related papers  (2026-09-01T07:42:45Z)
• HALOMI: Learning Humanoid Loco-Manipulation with Active Perception from Human Demonstrations [10.378782873235453]
  We present HALOMI, a framework for learning humanoid loco-manipulation with active perception from human demonstrations.<n>We propose a manifold-constrained controller that plans in a learned latent behavior manifold to enable precise and robust head-hand tracking in the world frame.<n>We validate HALOMI on a Unitree G1 humanoid robot with an actuated neck across five real-world tasks involving navigation, grasping, bimanual manipulation, whole-body coordination, and dynamic behaviors.
  arXiv  Detail & Related papers  (2026-06-17T07:33:37Z)
• LatentPilot: Scene-Aware Vision-and-Language Navigation by Dreaming Ahead with Latent Visual Reasoning [51.969318585152116]
  LatentPilot exploits future observations during training as a valuable data source to learn action-conditioned visual dynamics.<n>We propose a flywheel-style training mechanism that iteratively collects on-policy trajectories and retrains the model to better match the agent's behavior distribution.<n> Experiments on R2R-CE, RxR-CE, and R2R-PE benchmarks achieve new SOTA results, and real-robot tests across diverse environments demonstrate LatentPilot's superior understanding of environment-action dynamics in scene.
  arXiv  Detail & Related papers  (2026-03-31T02:21:59Z)
• Cortical Policy: A Dual-Stream View Transformer for Robotic Manipulation [57.28703268044067]
  We propose Cortical Policy, a novel dual-stream view transformer for robotic manipulation.<n>Our framework offers a fresh perspective for robotic manipulation and holds potential for broader application in vision-based robot control.
  arXiv  Detail & Related papers  (2026-03-22T04:18:54Z)
• Self-Correcting VLA: Online Action Refinement via Sparse World Imagination [55.982504915794514]
  We propose Self-Correcting VLA (SC-VLA), which achieve self-improvement by intrinsically guiding action refinement through sparse imagination.<n>SC-VLA achieve state-of-the-art performance, yielding the highest task throughput with 16% fewer steps and a 9% higher success rate than the best-performing baselines.
  arXiv  Detail & Related papers  (2026-02-25T06:58:06Z)
• Going with the Flow: Koopman Behavioral Models as Implicit Planners for Visuo-Motor Dexterity [1.6865201067073452]
  We introduce Unified Behavioral Models (UBMs), a framework that learns to represent dexterous skills as coupled dynamical systems.<n>Koopman-UBM is a first instantiation of UBMs that learns a unified representation in which the joint flow of latent visual and proprioceptive features is governed by a structured linear system.
  arXiv  Detail & Related papers  (2026-02-07T07:18:00Z)
• Learning Predictive Visuomotor Coordination [25.41320518416861]
  This work introduces a forecasting-based task for visuomotor modeling.<n>The goal is to predict head pose, gaze, and upper-body motion from egocentric visual and kinematic observations.<n>We extend a diffusion-based motion modeling framework that integrates egocentric vision and kinematic sequences.
  arXiv  Detail & Related papers  (2025-03-30T03:46:45Z)
• A Cognitive-Based Trajectory Prediction Approach for Autonomous Driving [21.130543517747995]
  This paper introduces the Human-Like Trajectory Prediction (H) model, which adopts a teacher-student knowledge distillation framework. The "teacher" model mimics the visual processing of the human brain, particularly the functions of the occipital and temporal lobes. The "student" model focuses on real-time interaction and decision-making, capturing essential perceptual cues for accurate prediction.
  arXiv  Detail & Related papers  (2024-02-29T15:22:26Z)
• Model-Based Reinforcement Learning with Isolated Imaginations [61.67183143982074]
  We propose Iso-Dream++, a model-based reinforcement learning approach. We perform policy optimization based on the decoupled latent imaginations. This enables long-horizon visuomotor control tasks to benefit from isolating mixed dynamics sources in the wild.
  arXiv  Detail & Related papers  (2023-03-27T02:55:56Z)
• Continual Visual Reinforcement Learning with A Life-Long World Model [55.05017177980985]
  We present a new continual learning approach for visual dynamics modeling.<n>We first introduce the life-long world model, which learns task-specific latent dynamics.<n>Then, we address the value estimation challenge for previous tasks with the exploratory-conservative behavior learning approach.
  arXiv  Detail & Related papers  (2023-03-12T05:08:03Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.