Technical advances in collaborative robots (cobots) are making them increasingly attractive to companies. However, many human operators are not trained to program complex machines. Instead, humans are used to communicating with each other on a task-based level rather than through specific instructions, as is common with machines. The gap between low-level instruction-based and high-level task-based communication leads to low values for usability scores of teach pendant programming. As a solution, we propose a task-based interaction concept that allows human operators to delegate a complex task to a machine without programming by specifying a task via triplets. The concept is based on task decomposition and a reasoning system using a cognitive architecture. The approach is evaluated in an industrial use case where mineral cast basins have to be sanded by a cobot in a crafts enterprise.

翻译：协作机器人（cobots）的技术进步正使其对企业越来越具吸引力。然而，许多人类操作员并未接受过编程复杂机器的培训。相反，人类习惯于基于任务层级相互沟通，而非像机器常见的那样通过具体指令进行交流。低层级指令式与高层级任务式交流之间的差距，导致示教器编程的可用性评分较低。作为解决方案，我们提出一种基于任务的交互概念，允许人类操作员通过三元组指定任务，无需编程即可将复杂任务委派给机器。该概念基于任务分解及使用认知架构的推理系统。该方法在工业用例中进行了评估——某手工艺企业需由协作机器人对矿物铸件底座进行打磨。