Coherent Methodology for Modelling and Design of Soft Material Robots – The Soft Material Robotics Toolbox (SMaRT)

在过去的几年里，不同的研究小组开发和研究了大量不同的软材料机器人(SMR)。由于其低刚性和高灵活性，SMR在处理物体或与其环境交互时具有优势。然而，在实际应用中开发和使用SMR是一项尚未克服的挑战。对于SMR的建模和控制，正在进行假设，这些假设简化了各自的任务，并且仅适用于非常具体的应用。尽管如此，为了利用SMR的优势，还需要开发一种更普遍的方法来将其应用。我们希望在这个项目中解决的一个空白是在SMR的建模、设计、运动规划和状态估计中整合更复杂的接触现象。到目前为止，建模的这一方面经常被忽视。然而，对于现实生活中的应用来说，考虑接触是必不可少的。由于人机交互或处理未知对象被称为SMR的最大优势，因此需要对交互进行研究。因此，与环境的接触，不能再被过分简单化或忽视。在第一个供资期间，开发了一个工具箱，其中包括支持SMR的开发、建模和优化的不同方法和途径，并以一个软气动执行器部分为例。在此前期工作的基础上，后续的研究项目旨在将工具箱扩展到现实生活中的应用程序。它将考虑多段执行器以及建模、设计、运动规划和动态控制方面的进一步发展。对于基于目标的模型设计过程，我们将研究和开发优化策略，考虑手头有特定任务的完整SMR，而不是关注单个软执行器的分离特性。对于这些优化问题，我们研究了在考虑障碍物的接触和利用的情况下，包含运动规划方案的可能性。因此，将开发和实现更复杂的、因此更现实的软材料系统的接触描述。关于现实生活中的应用，将研究结合障碍物利用的动作规划。为了动态地控制自由运动和估计系统的状态，同时又与环境接触，研究了基于模型的方法。总体目标是提供一个开源工具箱，促进软材料机器人的有针对性和面向应用的开发。

In the past years different research groups developed and investigated a large variety of different soft material robots (SMRs). Because of their low stiffness and high flexibility, SMRs are known to be advantageous when handling objects or interacting with their environment. However, development and use of SMRs for real life applications is a challenge that is not yet overcome. For modeling and control of SMRs assumptions are being made, that simplify the respective task and are only applicable for very specific applications. Still, to make use of the advantages of SMRs a more generalized approach to bring them into application needs to be developed. One gap that we want to tackle within this project is the integration of more complex contact phenomena in modeling, design, motion planning and state estimation of SMRs. This aspect of modeling has often been neglected until now. However, with regards to real life application it is essential to consider contact. Since human-robot interaction or handling unknown objects, is referred to as the greatest advantage of SMRs, interaction needs to be investigated. Consequently, contact with the environment, can no longer be oversimplified or neglected. During the first funding period a toolbox was developed that includes different methodologies and approaches that support the development, modeling and optimization of SMRs by the example of one soft pneumatic actuator segment. Based on this preliminary work, this subsequent research project aims to expand the toolbox with regards to real life applications. It will consider multi-segment actuators as well as further developments in modeling, design, motion planning and dynamic control.For a goal-oriented model based design process, we will research and develop optimization strategies that consider a complete SMR with specific tasks at hand rather than focusing on separated characteristics of a single soft actuator. For these optimization problems we research the possibility of including motion planning schemes also under consideration of contact and utilization of obstacles. Therefore, a more complex and thus more realistic contact description of soft material systems will be developed and implemented. With regards to real life applications motion planning in combination with the utilization of obstacles will be investigated. To dynamically control free motion and estimate the systems state, also in contact with the environment, model-based approaches are studied. The overall aim is to provide an open-source toolbox that facilitates a targeted and application-oriented development of soft material robots.