NRI: Small: Compliant Multifunctional Robotic Structures for Safety and Communication by Touch

这项研究的目的是通过使用新的兼容多功能结构，使机器人能够通过人类的触摸进行物理交流，从而更好地训练机器人。为了实现这一目标，导电聚合物阵列将被开发出来，形成一个类似于人类神经系统的系统，可以感知形状和力的分布。该传感器阵列将使用可扩展的增材制造工艺集成到复合泡沫结构中。为了支持模型的开发，并作为机器人平台上这些多功能结构的概念验证，模拟的协同机器人实验将使用机械臂与不同顺应性的物体相互作用进行。相关接触力学的实验细节将使用数字图像相关和传统视频成像实时量化。然后，通过使用新颖的奇异值分解方法求解非线性逆问题，将传感器阵列的输出与形状和力分布联系起来。研究结果将被记录和传播，实验将转化为旨在教育年轻女孩的STEM示范。这项研究将导致新的兼容，可扩展，传感结构，同时实时监测全局和局部形状，以及力分布。由于机器人还不存在兼容的多功能传感结构，因此设想所提出的工作将能够实现用于训练机器人的新的仿生控制原理。通过将机器人与人类的互动与环境中的其他物体的互动区分开来，这将大大提高协同机器人做出更安全互动和决策的能力。拟议的研究和教育的整合将培训新的机械工程师创造多功能产品，使新产品和新功能的现有产品在关键领域，如医疗保健。新的制造方法将使这些结构以具有成本竞争力的方式在美国制造，从而增加就业。

The objective of this research is to enable better training of robots by enabling them to physically communicate via human touch using new compliant multifunctional structures. To achieve this, arrays of conducting polymers will be developed to form a system similar to the human nervous system that can sense shape and forces distributions. This sensor array will be integrated into composite foam structures using a scalable additive manufacturing process. To support development of models and to serve as proof-of-concept for these multifunctional structures on robotic platforms, simulated co-robotics experiments will be conducted using a robotic arm interacting with objects of varying compliance. Experimental details of the associated contact mechanics will be quantified in real-time using Digital Image Correlation and conventional video imaging. Output from the sensor array will then be related to shape and force distributions by solving the nonlinear inverse problem using a novel Singular Value Decomposition method. Research results will be documented and disseminated, and the experiments will be converted to STEM demonstrations targeted at educating young girls.This research will lead to new compliant, scalable, sensing structures that simultaneously monitor in real-time both global and local shapes, as well as force distributions. Since compliant multifunctional sensing structures do not yet exist for robots, it is envisioned that the proposed work will enable realization of new bio-inspired control principles for training robots. This will significantly advance the ability to make safer interactions and decisions in co-robotics by differentiating robotic interactions with humans from other objects in their environment. The proposed integration of research and education will train new mechanical engineers to create multifunctional products that enable new products and new capabilities in existing products in critical areas such as healthcare. The new fabrication methods will enable these structures to be manufactured in the United States in a cost-competitive manner, increasing employment.