High-speed motion tracking and coupling for human-robot collaborative assembly tasks (HiSMoT)

用于人机协作装配任务的高速运动跟踪和耦合 (HiSMoT)

• 批准号: 500490184
• 负责人: Professor Dr.-Ing. Bernd Kuhlenkötter
• 金额: --
• 依托单位: Lehrstuhl für Fertigungsautomatisierung und Montage (FAMS)
• 依托单位国家: 德国
• 项目类别: Research Grants
• 资助国家: 德国
• 项目状态: 未结题
• 来源: https://gepris.dfg.de/gepris/projekt/500490184?language=en
• 关键词: speed; motion; tracking; coupling; human

Human-robot collaboration (HRC) combines the capabilities of humans and robots in order to create a more inclusive and human-centered future in the production industry. The capabilities of HRC have been investigated extensively for manufacturing processes, mainly in laboratory environments. However, they are not widely applied in production systems. Performance and safety can be a primary challenge that limits collaboration efficiency, particularly for motions at higher speeds. The collaborative handling of an object is one of the current challenges facing HRC's real collaboration capabilities. In this context, how robots and humans should behave to handle an object and anticipate motion for mutual care is still unclear. One approach is to create robot motions that avoid risks to human body parts caused by high speeds of the robot tool center point (TCP), which might result from reactions to human motions. Coupling motion models for such approaches is difficult because unlike industrial robots, human motion is not easily modeled using geometric or analytical formulations such as rigid body dynamics. Instead, data-driven models for human motion generation are employed in various research and industrial applications such as sports, health, film, ergonomics and production. This research project investigates real-time capable approaches to coupling data driven human motion models with robot motion models in HRC applications. The project concentrates on the collaborative handling of rigid parts. In real time modeling, high speed motions are considered. A central research question is whether and in what way human and robot motion models can be coupled and implemented using motion capture-driven approaches. In this context, latent space control approaches are investigated for the capability to follow the high mutual support and anticipation principle. With the evolving concept of a personalized production system, speed, accuracy, agility and controllability are key parameters. In this regard, HRC may have the potential to narrow a gap in the automation of product assembly systems.

人机协作（HRC）结合了人类和机器人的能力，以便在生产行业中创造更具包容性和以人为本的未来。HRC的能力已被广泛研究的制造过程中，主要是在实验室环境。然而，它们在生产系统中并没有广泛应用。性能和安全性可能是限制协作效率的主要挑战，特别是对于高速运动。对象的协作处理是HRC的真实的协作能力所面临的当前挑战之一。在这种情况下，机器人和人类应该如何处理物体并预测相互关心的运动仍然不清楚。一种方法是创建机器人运动，以避免由机器人工具中心点（TCP）的高速引起的对人体部位的风险，这可能是由于对人体运动的反应。这种方法的耦合运动模型是困难的，因为不像工业机器人，人体运动不容易使用几何或分析公式，如刚体动力学建模。相反，用于人体运动生成的数据驱动模型被用于各种研究和工业应用，如体育，健康，电影，人体工程学和生产。本研究项目探讨了在HRC应用中将数据驱动的人体运动模型与机器人运动模型耦合的实时方法。该项目专注于刚性零件的协同处理。在真实的时间建模中，考虑高速运动。一个中心的研究问题是人类和机器人的运动模型是否以及以何种方式可以耦合和实现运动捕捉驱动的方法。在这种背景下，潜在的空间控制方法的能力，遵循高相互支持和预期的原则进行了研究。随着个性化生产系统概念的不断发展，速度、准确性、灵活性和可控性成为关键参数。在这方面，HRC可能有潜力缩小产品装配系统自动化方面的差距。