Symbolic and Qualitative Reasoning for Error Recovery in Robot Programs

这项研究探索了机器人装配任务的错误处理自动化的可能性。处理因错误、故障或其他计划外情况而产生的不确定性的能力对于机器和机器人的有效性至关重要。这种能力取决于能够访问有关任务、对象、工具和工作环境的信息的机器。这个项目的中心问题是，机器需要哪些信息来检测和处理意外情况。这涉及确定所需信息的特征、自动获取信息、决定信息的适当表示以供实时使用、区分错误条件和正常事件、跟踪错误原因以及生成恢复程序等问题。

This research explores the possibilities of automation of error-handling for robot assembly tasks. The ability to cope with uncertainties due to errors, failures, or other unplanned situations is critical to the effectiveness of machines and robots. That capability depends on the machines having access to information about tasks, objects, tools, and the work environment. This project centers on the question of what information needs to be available to a machine to detect and handle unexpected situations. This involves issues of characterizing the required information, automating its acquisition, deciding on appropriate representations of the information for real-time use, distinguishing an error condition from normal events, tracing causes of errors, and generating recovery procedures.