Analysis of self-adaptive robotic mechanisms

本申请中提出的研究计划旨在开发新的自适应机器人系统的新架构，以及探索这些系统的运动学的有趣特性。特别是，该研究计划将集中在1）使用连杆驱动传动装置的自适应机制的设计两个新的应用和2）触觉探索模型的开发。在过去的几十年里，机电一体化系统已经提出利用自适应机制。这些机构的特征在于具有将控制任务的一部分委托给系统的机械布局而不是电子板的性质。自适应机构最著名的应用是欠驱动抓取，其中致动器的闭合运动被机械地分配到机器人手的几个自由度（DOF）。这导致机器人手在没有任何传感器或专用控制规则的情况下机械地适应所抓取物体的形状。自适应机制只是最近才被认为是在机器人，因为仍然是最常见的趋势一直期望，不断增长的计算设备的能力将允许轻松地控制所有系统，尽管其固有的复杂性，具有很高的效率。然而，这一预测只是部分实现了现代数字系统的计算能力的增加并不总是能够克服所有这些缺点。在过去的十年中，申请人已经深入研究了欠驱动机器人手，并且该研究工作的结果导致发现了数百种新的机械架构和基于驱动连杆运动学的分析来合成它们的方法。

The research program proposed in this application aims at developing new architectures of self-adaptive robotic systems for novel applications as well as exploring intriguing properties in the kinematics of these systems. In particular, the research program will focus on 1) the design of self-adaptive mechanisms using linkage-driven transmissions for two novel applications and 2) the development of haptic exploration models. Over the past decades, mechatronic systems have been proposed taking advantage of self-adaptive mechanisms. These mechanisms are characterized by the property of having a part of the control tasks delegated to the mechanical layout of the system rather than to an electronic board. The most well-known application of self-adaptive mechanisms is underactuated grasping where the closing motion of an actuator is mechanically distributed to the several degrees of freedom (DOF) of a robotic hand. This results in a robotic hand which adapts mechanically to the shape of the seized objects without any sensors or dedicated control laws. Self-adaptive mechanisms have only recently been considered in robotics since the still most common trend has been to expect that the growing power of computational devices would allow to easily control all systems with great efficiency despite their inherent complexity. However, this prediction has only been partially fulfilled as the increase of the computation capabilities of modern digital systems has not always been able to overcome all these drawbacks. Underactuated robotic hands have been intensively studied by the applicant during the last decade and the results of this research effort lead to the discovery of hundreds of new mechanical architectures and a methodology to synthesize them based of the analysis of the driven linkage kinematics.