---
格式版本: 2
标题: "Real-Time Dynamics-Based Torque-Sampling MPPI for Compliant and Force Aware Manipulation"
原文链接: "https://arxiv.org/abs/2609.02020"
发布日期: "2026-09-02"
发布时间校准状态: "found"
发布时间需复核: "否"
发布时间来源: "rule:local:strict_original_body"
发布时间证据: "**\\[v1\\]** Wed, 2 Sep 2026 02:46:57 UTC (481 KB)"
发布时间校准原因: "规则确认唯一严格发布时间，来源 local:strict_original_body"
发布时间校准置信度: "high"
发布时间候选数量: 18
发布时间严格候选数量: 6
发布时间原页读取状态: "source template page reused from URL open"
发布时间未找到原因: ""
发布时间校准时间: "2026-09-04T01:34:53+08:00"
发布时间仲裁状态: "skipped"
发布时间仲裁尝试次数: 0
发布时间仲裁耗时毫秒: 0
发现时间: "2026-09-04T01:32:14+08:00"
入库时间: "2026-09-03T17:34:53.728Z"
来源平台: "arXiv 学术论文搜索"
搜索渠道: "source_template"
搜索词: "https://arxiv.org/search/?query=GPU&searchtype=all"
匹配关键词:
  - "GPU"
  - "performance"
相关厂家:
  []
相关专家:
  []
内容类型: "网页"
抓取工具: "Free Fetch + Defuddle"
清洗工具: "Defuddle Markdown + Defuddle/Readability 正文提取"
原始附件:
  []
AI优质: "否"
AI打分: 8
AI分档: "非优质"
AI质检状态: "不通过"
AI打分理由: "该资料为arXiv机器人操纵论文，主题为MPPI控制与力觉交互，与超节点/AI Rack/机柜级AI基础设施完全无关。仅提及GPU并行化用于控制，非数据中心硬件。"
AI质检模型: "zj-deepseek-v4-flash"
AI质检时间: "2026-09-04T01:35:29+08:00"
AI主题相关性: 0
AI来源权威性: 8
AI新颖性: 0
AI技术细节: 0
AI商业部署信号: 0
AI完整性: 0
AI摘要: "该研究提出一种基于模型预测路径积分（MPPI）的任务空间控制框架，通过实时求解刚体动力学并施加安全约束，实现机械臂的柔顺与力感知操作；采用扭矩采样架构配合GPU并行，控制更新率超过166Hz，并在7自由度机械臂上完成实验验证。"
AI摘要模型: "ali-deepseek-v4-flash"
AI摘要时间: "2026-09-04T00:18:03.236Z"
采集批次: "2026年9月3日22点43分34秒"
采集批次ID: "20260903-224334-406"
去重键: "https://arxiv.org/abs/2609.02020"
---

## Computer Science > Robotics

## Title:Real-Time Dynamics-Based Torque-Sampling MPPI for Compliant and Force Aware Manipulation

Authors:[Euncheol Im](https://arxiv.org/search/cs?searchtype=author&query=Im,+E), [Taehyun Kim](https://arxiv.org/search/cs?searchtype=author&query=Kim,+T), [Yonghwan Oh](https://arxiv.org/search/cs?searchtype=author&query=Oh,+Y), [Myotaeg Lim](https://arxiv.org/search/cs?searchtype=author&query=Lim,+M), [Yisoo Lee](https://arxiv.org/search/cs?searchtype=author&query=Lee,+Y)

[View PDF](https://arxiv.org/pdf/2609.02020) [HTML (experimental)](https://arxiv.org/html/2609.02020v1)

> Abstract:This study proposes a novel Model Predictive Path Integral (MPPI)-based task-space control framework. The proposed framework explicitly solves rigid-body dynamics within a real-time MPC formulation and enforces safety constraints, enabling accurate motion and force control that yields compliant behaviors for safe and effective physical interaction of robotic manipulators in unstructured environments. By leveraging MPPI, the proposed framework efficiently handles nonlinear dynamics that are difficult to solve with conventional MPC approaches in real-time. Furthermore, we develop a torque-sampling-based control architecture that enables efficient exploitation of GPU-based parallelization, resulting in effective compliant and force-aware behaviors. As a result, the proposed framework achieves a solver update rate of over 166 Hz with a 0.18 s prediction horizon, and its performance is validated through real-world experiments on a 7-DoF manipulator.

| Comments: |  |
| --- | --- |
| Subjects: | Robotics (cs.RO) |
| Cite as: | [arXiv:2609.02020](https://arxiv.org/abs/2609.02020) \[cs.RO\] |
|  | (or [arXiv:2609.02020v1](https://arxiv.org/abs/2609.02020v1) \[cs.RO\] for this version) |
|  | [https://doi.org/10.48550/arXiv.2609.02020](https://doi.org/10.48550/arXiv.2609.02020) |

## Submission history

From: Euncheol Im \[[view email](https://arxiv.org/show-email/de3838e9/2609.02020)\]  
**\[v1\]** Wed, 2 Sep 2026 02:46:57 UTC (481 KB)

[Which authors of this paper are endorsers?](https://arxiv.org/auth/show-endorsers/2609.02020) | Disable MathJax ([What is MathJax?](https://info.arxiv.org/help/mathjax.html))