以可微方式参数化手指滚动以及手指与物体之间的接触,对于将灵巧操作表述为轨迹优化问题十分重要。以往方法通常假设机器人和物体具有简化几何形状,或者没有显式建模手指滚动;与之不同,我们提出一种方法,通过同时考虑机器人和物体的非平凡几何形状,进一步扩展灵巧操作的能力。通过将物体的 Signed Distance Field(SDF,有符号距离场)与一种采样方法相结合,我们的方法以可微方式估计接触及与滚动相关的变量,并将这些变量纳入轨迹优化框架。该表述自然允许手指滚动行为涌现,使机器人能够局部调整接触点。为评估我们的方法,我们引入了一个基准,其中包含具有挑战性的多指灵巧操作任务,例如转动螺丝刀和手内重新定向。在达到期望物体构型以及避免物体掉落方面,我们的方法优于基线方法。我们还将该方法成功应用于真实世界中的螺丝刀转动任务和长方体对齐任务,证明了其对 sim2real 差距的鲁棒性。
英文摘要:
Parameterizing finger rolling and finger-object contacts in a differentiable manner is important for formulating dexterous manipulation as a trajectory optimization problem. In contrast to previous methods which often assume simplified geometries of the robot and object or do not explicitly model finger rolling, we propose a method to further extend the capabilities of dexterous manipulation by accounting for non-trivial geometries of both the robot and the object. By integrating the object's Signed Distance Field (SDF) with a sampling method, our method estimates contact and rolling-related variables in a differentiable manner and includes those in a trajectory optimization framework. This formulation naturally allows for the emergence of finger-rolling behaviors, enabling the robot to locally adjust the contact points. To evaluate our method, we introduce a benchmark featuring challenging multi-finger dexterous manipulation tasks, such as screwdriver turning and in-hand reorientation. Our method outperforms baselines in terms of achieving desired object configurations and avoiding dropping the object. We also successfully apply our method to a real-world screwdriver turning task and a cuboid alignment task, demonstrating its robustness to the sim2real gap.
Compliant In-hand Rolling Manipulation Using Tactile Sensing
Center for Robotics and Biosystems and the Department of Mechanical Engineering, Northwestern University(机器人与生物系统中心和机械工程系,西北大学);Northwestern Institute on Complex Systems (NICO)(复杂系统研究所(NICO))
E-TRoll: Tactile Sensing and Classification via A Simple Robotic Gripper for Extended Rolling Manipulations
Imperial College London(伦敦帝国理工学院)
机器人触觉感知提供了一种在视觉失效的情况下识别物体及其属性的方法。以前的工作在机器人操作的触觉感知经常集中在探索性的程序(EP)。然而,同样受人类启发的手动操作技术可以在EP的一小部分时间内收集丰富的数据。我们提出了一个简单的3自由度机器人手的设计,优化对象滚动任务通过一个可变宽度的手掌和相关的控制系统。该系统响应于对象行为动态地调整手指基部之间的距离。与固定的手指基座相比,该技术显著增加了在单个滚动运动期间暴露于安装在手指上的触觉阵列的对象的面积(对于具有30毫米直径的圆柱体观察到超过60%的增加)。此外,本文提出了一种时空数据集的特征提取算法,重点是目标角点的识别、分析和紧凑表示。该技术将每个数据样本的维度从10 x 1500时间序列数据大幅降低到80个特征,并通过主成分分析(PCA)进一步降低到22个成分。一个集成子空间k-最近邻(KNN)分类模型进行了训练,90个观察滚动三个不同的几何对象,在三倍交叉验证准确率为95.6%的对象形状识别。
Grasp, Slide, Roll: Comparative Analysis of Contact Modes for Tactile-Based Shape Reconstruction
Amazon Fulfillment Technologies & Robotics(亚马逊履行技术与机器人);Robotics Institute at Carnegie Mellon University(卡内基梅隆大学机器人研究所);Siebel School of Computing and Data Science, University of Illinois at Urbana-Champaign(伊利诺伊大学厄巴纳-香槟分校Siebel计算与数据科学学院)
MOBIUS: A Multi-Modal Bipedal Robot that can Walk, Crawl, Climb, and Roll
Mitsubishi Electric Research Laboratories(三菱电机研究实验室);Robotic Systems Lab(机器人系统实验室);Robotics and Mechanisms Laboratory(机器人与机构实验室);Department of Mechanical and Aerospace Engineering, University of California, Los Angeles(加州大学洛杉矶分校机械与航空航天工程系)