Collaborative robots: novel architectures and interaction paradigms

这项研究的长期目标是开发机器人机械系统的新概念，这将导致轻量级，敏捷，有效和响应的机器人具有与人类相似的机械能力，包括与人类及其环境进行物理交互的能力。为了将机器人融入以人为中心的环境，如工厂、医院、家庭和其他环境，需要这种能力。物理人机交互（physical human-robot interaction，pHRI）旨在利用机器人和人类的互补能力，让它们通过直接交互一起工作。pHRI中的关键挑战之一是提供对于人类用户来说直观的交互。在这项研究中，建议开发用户兼容的机器人，以产生一个非常直观的互动。换句话说，这项研究的目的是开发在接触力作用下表现得与人类相似的机器人。首先，将重新审视机器人、人类用户和环境之间相互作用的机械建模。基于对由此获得的数学模型的分析，将基于机器人和环境的特性来识别各种潜在的交互模态。在这项工作中，反向驱动能力-即，人类用户通过在机器人的连杆上施加力来直接移动机器人的能力将作为获得直观交互的手段而受到青睐。与反向驱动能力相关的挑战之一是所需电机的尺寸和重量。为了应对这一挑战，将研究不同的机械架构，例如并联机构或冗余执行器。一个高反应性的协作机器人的原型将设计和建造的基础上的结果。将进行实验，以验证系统的能力，提供一个直观的交互与人类用户和环境。 另一个研究目标是开发协作机器人手臂，其中末端执行器的位置和方向的本体感受测量与致动分离，这为超轻量化设计开辟了道路。还将建造一个原型来演示这一概念。超轻型机器人构成了pHRI的一个非常有前途的途径，因为它们大大降低了用户受伤的风险。这里提出的研究有可能大大有助于理解pHRI的机制，特别是当机器人和用户需要与环境交互时。此外，使用新方法的高效机器人机制将被设计和实验测试。这项工作中开发的新技术将引起工业界的极大兴趣，并有可能为人类友好型机器人的发展做出重大贡献。

The long-term objective of this research is to develop novel concepts of robotic mechanical systems that will lead to lightweight, agile, effective and responsive robots having mechanical capabilities that are similar to those of human beings, including abilities to physically interact with humans and with their environment. Such capabilities are required in order to integrate robots in human-centred environments such as factories, hospitals, households and others. Physical human-robot interaction (pHRI) aims at taking advantage of the complementary capabilities of robots and humans by letting them work together through direct interaction. One of the key challenges in pHRI is to provide an interaction that is intuitive for a human user. In this research, it is proposed to develop user compliant robots in order to produce a very intuitive interaction. In other words, this research aims to develop robots that behave similarly to humans when contact forces are applied to them. First, the mechanical modelling of the interaction between a robot, a human user and the environment will be revisited. Based on the analysis of the mathematical models thus obtained, a variety of potential interaction modalities will be identified based on the characteristics of the robot and of the environment. In this work, backdrivability - i.e., the ability for a human user to directly move a robot by applying forces on its links - will be favoured as a means of obtaining an intuitive interaction. One of the challenges associated with backdrivability is the size and weight of the motors that are required. In order to address this challenge, different mechanical architectures will be investigated such as, for instance, parallel mechanisms or redundant actuators. A prototype of a highly reactive collaborative robot will be designed and built based on the results. Experiments will be conducted in order to validate the capabilities of the system to provide an intuitive interaction with a human user and with the environment. Another research objective is the development of collaborative robot arms in which the proprioceptive measurement of the position and orientation of the end-effector is dissociated from the actuation, which opens the avenue to ultra-lightweight designs. A prototype will also be built to demonstrate this concept. Ultra-lightweight robots constitute a very promising avenue for pHRI because they considerably reduce the risk of injury to users. The research proposed here has the potential to greatly contribute to the understanding of the mechanics of pHRI, especially when robots and users need to interact with the environment. Also, highly effective robotic mechanisms using novel approaches will be designed and tested experimentally. The novel technologies developed in this work will be of great interest to industry and will have the potential to contribute significantly to the development of human-friendly robots.