Model-based design of granulate-based grippers for the form-variable handling in automated production

日益增加的部件复杂性和变体多样性给自动化生产带来了重大挑战，特别是在部件处理的灵活性方面。集装箱中高度不同的组件的无序提供导致了这些组件的自动分离(分拣)的一些技术挑战。由于到目前为止，标准的末端执行器通常用于垃圾箱的拾取，越来越多的工作被投入到基于传感器的单个部件的定位上，以确保稳健的分离过程。然而，如何通过足够普遍和强大的抓取技术来减少对适应性网络规划的需求仍然是未知的。这是使用可变形状抓取技术有效实施料箱拣选过程所需的基础研究的基础。因此，在本项目中，将详细阐述在IWF设计的装卸系统的基于模型的设计方法。处理系统的有效特定流程设计、配置和控制将使仓储挑选流程的可靠实施成为可能。研究方法是在自己前期工作的基础上，提出了一种基于颗粒的变形状、变面积抓取技术，并提出了基于静态特征图的设计方法。虽然这项技术能够显著提高使用同一工具抓取各种对象的灵活性，但同时夹爪设计的复杂性也增加了。在以前的实验设计方法中，为了能够控制设计的复杂性，有必要简化所考虑的因果关系。必须承认的是，尽管所产生的解可以以较高的实验成本实现，但它们可能远不是最优解。有限的溶液质量和低的设计效率促使人们研究基于颗粒的处理技术的新的设计方法。本项目将开发的基于模型的设计方法使人们能够对将要投资的设计工作做出积极的决定。根据夹爪设计(使用经验元模型和空间解析的数值离散元模拟)的预期努力依赖的结果质量，该方法建议了一种合适的下游工艺规划策略，该策略能够实现稳健的料箱挑选过程。在成分谱的基础上，开发了设计方法，并进行了实验验证和推广，以适用于类似的问题。其结果是基于开发的设计方法的可转移的参考程序，该方法考虑了综合的设计努力和质量。

Increasing component complexity and variant diversity lead to major challenges in automated production, especially in the flexibilization of parts handling. The disordered provision of highly variant components in containers results in a number of technical challenges for the automated separation of these components (bin picking). Since standardized end effectors have usually been used in bin picking up to now, more and more effort is being put into sensor-based localization of individual components to ensure a robust separation process. However, it is still unknown how the need for adaptive web planning can be reduced by sufficiently universal and robust gripping technology. This is the basis for the fundamental research required for the efficient implementation of bin picking processes using variable shape gripping technology. In this project, a model-based design method for a handling system designed at IWF will therefore be elaborated. The efficient process-specific design, configuration and control of the handling system is to be enabled for the robust implementation of bin picking processes. The research approach is based on own preliminary work, in which a granulate-based handling technology for variable shape and area gripping was proposed and a design methodology based on static characteristic maps was developed. While this technology enables a significant increase in flexibility with regard to gripping a wide variety of objects with the same tool, the complexity of gripper design increases at the same time. With previous experimental design approaches, it was necessary to simplify the considered cause-and-effect relationships in order to be able to control the design complexity. It must be accepted that although the solutions generated can be realized at high experimental cost, they may be far from an optimal solution. The limited solution quality and the low design efficiency motivate research into new design methods for granule-based handling technology. The model-based design methodology to be developed in this project enables an active decision to be made on the design effort to be invested. Depending on the expected effort-dependent result quality of the gripper design (using an empirical meta model and spatially resolved numerical discrete element simulation), this methodology suggests a suitable downstream process planning strategy that enables a robust bin picking process. On the basis of a component spectrum, the design methodology is developed, experimentally validated and generalized for transferability to similar problems. The result is a transferable reference procedure based on the developed design methodology, which takes into account the combined design effort and quality.