The Netherlands Organisation for Applied Scientific Research-TNO, 2316 ZL Leiden, The Netherlands.
Department of Sciences and Methods of Engineering, University of Modena and Reggio Emilia, 42122 Reggio Emilia, Italy.
Sensors (Basel). 2022 Jun 29;22(13):4901. doi: 10.3390/s22134901.
Effective task scheduling in human-robot collaboration (HRC) scenarios is one of the great challenges of collaborative robotics. The shared workspace inside an industrial setting brings a lot of uncertainties that cannot be foreseen. A prior offline task scheduling strategy is ineffective in dealing with these uncertainties. In this paper, a novel online framework to achieve a resilient and reliable task schedule is presented. The framework can deal with deviations that occur during operation, different operator skills, error by the human or robot, and substitution of actors, while maintaining an efficient schedule by promoting parallel human-robot work. First, the collaborative job and the possible deviations are represented by AND/OR graphs. Subsequently, the proposed architecture chooses the most suitable path to improve the collaboration. If some failures occur, the AND/OR graph is adapted locally, allowing the collaboration to be completed. The framework is validated in an industrial assembly scenario with a Franka Emika Panda collaborative robot.
在人机协作 (HRC) 场景中，有效的任务调度是协作机器人面临的重大挑战之一。工业环境中的共享工作空间带来了许多无法预见的不确定性。预先离线的任务调度策略在应对这些不确定性方面效果不佳。本文提出了一种新的在线框架，以实现具有弹性和可靠性的任务调度。该框架可以处理操作过程中出现的偏差、不同操作员的技能水平、人机错误以及角色替换等情况，同时通过促进人机并行工作来保持高效的调度。首先，协作作业和可能的偏差由 AND/OR 图表示。随后，所提出的架构选择最合适的路径来改善协作。如果发生某些故障，AND/OR 图会在本地进行调整，从而允许协作完成。该框架在一个带有 Franka Emika Panda 协作机器人的工业装配场景中得到了验证。
