郑刚，男，现为法国国家信息自动化研究所一级研究员、博导，IEEE高级会员，法国动态系统建模、分析与决策委员会会员，法国里尔大学中法联合实验室副主任。曾任法国国家学术委员会控制专委会副主席，法国北方自动化和人机智能系统委员会副主席, 以及国际自动控制联合会(IFAC) TC 9.2技术委员会副主席，其主要研究兴趣包括复杂动态系统的观测与控制、及其在机器人中的相关应用。目前在Springer出版专著1本，在国际会议及期刊上发表学术论文160余篇，包括约20篇Automatica及IEEE TAC，参与并主持多项法国国家自然科学基金项目。
讲座地点：腾讯会议直播（ID：368 227 364 链接：https://meeting.tencent.com/dm/Vlyebpch4RFw）
摘要: Recently soft robotics has rapidly become a novel and promising area of research with many designs and applications due to their flexible and compliant structure. However, it is more difficult to derive the nonlinear dynamic model of such soft robots. In this talk, two different modeling techniques will be presented. The first method is based on Cosserat rod theory, where micro structure has been imposed for the continuum medium to facilitate the modeling of soft slender type of robot. By applying Newton-Euler approach, the differential kinematics and dynamics of the soft manipulator can be formulated as a set of highly nonlinear partial differential equations via the classic Cosserat rod theory. Then a discrete modeling technique, named piecewise linear strain, is proposed to solve the deduced PDEs, based on which the associated analytic models are obtained. The second modeling approach is the so-called finite element method to model soft robot with arbitrary shape, where the particles are only equipped with position vectors.
讲座地点：腾讯会议直播（ID：394 451 613 链接：https://meeting.tencent.com/dm/bbqGLSdoY4dk）
摘要: Nowadays, the process of designing soft robots is still governed by trial-and-error or bio-inspired notions. Given a particular soft robot’s configuration, evaluating its reachable workspace is still an open subject, and is very essential for other soft robotics’ main scientific challenges, such as control, trajectory planning, and design optimization. For this topic, three different techniques will be presented in this talk. The first one is an optimization-based approach that consists of mapping the exterior boundaries of the workspace. This method can successfully reduce the complexity of the workspace estimation compared to the forward approach, but cannot provide interior knowledge to the workspace. The second approach is the interval analysis technique which consists of exploring all feasible configurations in the workspace of soft robots. However, this method is relatively exhaustive since it consists of exploring the whole feasible configurations in the workspace. To reduce the computation complexity, an alternative methodology to determine all boundaries of soft robots’ workspaces, named as the continuation approach, will be presented.
讲座地点：腾讯会议直播（ID：168 887 107 链接：https://meeting.tencent.com/dm/V5Hjf5Ja8wfg）
摘要: Unlike rigid robotics, elastic deformation of soft robot results in infinite degrees-of-freedom motions. Therefore, the control theory developed for rigid robot was poorly applicable in this case. Moreover, numerical models used for modeling deformation were not particularly well adapted for feedback control: too slow to compute and too difficult to analyze (because of the lack of analytical form). In the literature, a lot of researches are using model-free approaches for control, where PID controller is the simplest and most used one to control soft robots. However, such a controller can only achieve local and slow-motion tracking. Another method is driven by empirical data and applies machine learning techniques to design controllers. Such an approach mainly works for static models (thus cannot realize fast motion control), and the learned model varies case by case (thus cannot treat the unpredictable external disturbance). In this talk, we present how to use precise deformation modeling methods (Cosserat and FEM) for the control, since it allows us to take into account the physics (and even the multi physics) of the robot and the interaction with the environment.