# 非抓取投掷：强化学习视角

> 原文链接：https://robotics.hellobike.com/papers/2026-09-01-2609-00771
> 频道：papers | 更新日期：2026-09-01 | 标签：arxiv, cs.RO
> 核心摘要：提出强化学习非抓取投掷方法，直接优化关节轨迹，并在仿真和UR5e真机零样本部署中验证。

# 非抓取投掷：强化学习视角

::: v-pre
> 原文链接：[Non-Prehensile Throwing: A Reinforcement Learning Perspective](https://arxiv.org/abs/2609.00771)（arXiv 2609.00771 · 发布日期：2026-09-01）

**Non-Prehensile Throwing: A Reinforcement Learning Perspective**

提出强化学习非抓取投掷方法，直接优化关节轨迹，并在仿真和UR5e真机零样本部署中验证。

## 核心要点

- 面向大、重或可变形物体，非抓取投掷可扩展传统抓放之外的工作空间。
- 方法利用滑动和滚动接触模式，无需解析接触模型，直接优化关节空间轨迹。
- 仿真中在数千配置上达到99%成功率，并可泛化到未见物体。
- 真机零样本部署在UR5e上，投掷多类物体到远距或高处目标，成功率97%。

**验证范围与局限**：验证范围包括仿真数千配置和UR5e真机非抓取投掷任务，真机运行接近物理极限。

## 英文摘要 (Abstract)

Robotic throwing enables fast object transport and extends a robot's reachable workspace beyond traditional pick-and-place. While prehensile (grasp-based) throwing works well for graspable items, non-prehensile (grasp-free) throwing is better suited for large, heavy, and/or deformable objects. Existing approaches rely on model-based optimization with simplified contact models (e.g., dynamic grasping) and low-dimensional trajectory parameterizations, which limit solution quality and reachable workspace. We propose a reinforcement learning approach that additionally leverages sliding and rolling contact modes and directly optimizes joint-space trajectories without analytical contact models or custom parameterizations. The Markov Decision Process (MDP) is formulated as a dynamical system that evolves the robot's joint state conditioned on the throwing target, object model, and initial configuration. Joint-jerk trajectories are planned offline at a low control rate and upsampled into smooth, high-rate velocity commands for deployment. For sim-to-real transfer, we minimize the robot-dynamics gap through minimum-jerk system identification and train uncertainty-aware policies to mitigate object-modeling errors, particularly sensitivity to dynamic friction. In simulation, the policy achieves 99% success across thousands of configurations and generalizes to unseen objects. Sensitivity analysis shows robustness to mass uncertainty but high sensitivity to dynamic friction, consistent with the sliding-based release mechanism. Deployed zero-shot on a UR5e operating near its physical limits (5 m/s end-effector velocity), our method throws diverse objects including heavy (790 g) and large (20x20x28 cm) items to targets up to 350 cm distance or 180 cm elevation, achieving a 97% real-world success rate.

作者：Abdullah Mustafa, Ryo Hanai, Ixchel G. Ramírez-Alpizar, Floris Erich, Ryoichi Nakajo, Yukiyasu Domae, Tetsuya Ogata
:::

## 原始参考来源 (Sources)

- [arXiv 2609.00771：Non-Prehensile Throwing: A Reinforcement Learning Perspective](https://arxiv.org/abs/2609.00771)