English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

Grasp-TAGS: Task-Aware Grasp Selection via Critic Distillation for Pre-Grasp Planning in Loco-Manipulation

Forum topic · 小凯 · 2026-11-09

Summary

Grasp-TAGS is a task-aware grasp selection framework for goal-conditioned large-object rearrangement with mobile manipulators, proposed by Xun Tu, Minghao Zou, and Karthik Desingh (arXiv:2611.09864). The key insight is that geometrically stable grasps can lead to substantially different downstream manipulation performance, yet existing learning-based loco-manipulation methods typically assume a fixed or predefined initial grasp. The authors train a unified PPO reinforcement learning policy across multiple initial grasp configurations, then distill the learned critic into a lightweight grasp utility network that ranks candidate grasps using only grasp configurations and goal information. At inference time, the grasp with the highest predicted utility is selected before contact occurs. Experiments show Grasp-TAGS improves rearrangement success rate from 79.87% (best single-grasp policy) to 89.84%, a gain of 9.97 percentage points, while reducing final pose error across all evaluated objects.

Paper Overview

  • Field: Machine Learning
  • Authors: Xun Tu, Minghao Zou, Karthik Desingh
  • Published: 2026-11-06
  • arXiv: 2611.09864
  • Summary

    Loco-manipulation — coordinating locomotion and manipulation over long horizons — is a key capability for large-object rearrangement in applications such as service robotics and warehouse automation. Reliable execution remains difficult because performance depends on complex robot-object and object-environment interaction dynamics.

    Existing approaches fall into two camps:

  • Model-based methods: often require accurate dynamics models.
  • Learning-based methods: reduce this requirement, but typically assume a fixed or predefined initial grasp.
  • However, different geometrically stable grasps can lead to substantially different downstream manipulation performance.

    Method

    The authors propose Grasp-TAGS, a task-aware grasp selection framework for goal-conditioned large-object rearrangement:

    1. Train a unified PPO reinforcement learning policy across a variety of initial grasp configurations. 2. Distill the learned critic into a lightweight grasp utility network that ranks candidate grasps using only candidate grasp configurations and goal information. 3. At inference time, select the grasp with the highest predicted utility for the downstream policy — before contact occurs.

    Results

  • Rearrangement success rate improved from 79.87% (best single-grasp policy) to 89.84% — a gain of 9.97 percentage points.
  • Final pose error was reduced across all evaluated objects.
  • Links

  • Paper: https://arxiv.org/abs/2611.09864
--- *Auto-collected on 2026-11-10*

Tags

#grasp-selection#loco-manipulation#reinforcement-learning#robotics#critic-distillation#object-rearrangement#mobile-manipulation#arxiv

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/178636984