Event
Proposes VQ-WM latent feedback motion planning framework
1 reports · 1 sources
Xuezhou Xu et al. propose the Vector Quantized World Model (VQ-WM) latent feedback motion planning framework for long-horizon motion planning in robotic manipulation. The framework combines discrete long-range search with continuous visual feedback control: a discrete latent state predictor models transitions between vector-quantized latent states, enabling the planner to search for shortest paths over long distances; a continuous learned controller then reaches the planned states step by step via visual feedback, correcting deviations and mitigating compounding errors. The paper reports evaluation on a real-robot Tower of Hanoi manipulation task, achieving a 100% success rate using only short-horizon data; it also adopts a novel first-hit-time estimator as a quasi-metric that aligns the planning abstraction with the controller's capabilities.
Report timeline
- VQ-WM: A Latent Feedback Motion Planning Framework for Robot Manipulation
The paper proposes Vector-Quantized World Model (VQ-WM), a latent feedback motion planning framework that combines discrete long-horizon search with continuous visual feedback control: a discrete latent state predictor models transitions between vector-quantized latent states, allowing the planner to search for the shortest path over long distances, while a continuous learned controller then reaches the planned states step by step through visual feedback to correct deviations and mitigate compounding errors.
arXiv cs.RO ·