arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2209.14045cs.ROcs.AI
收藏

事件B中行为树的安全性

Verifying Safety of Behaviour Trees in Event-B

KTH Royal Institute of Technology(KTH皇家理工学院)

Matteo Tadiello, Elena Troubitsyna

摘要:

行为树(BT)在机器人社区中越来越受欢迎。BT工具非常适合决策应用程序,允许机器人执行复杂的行为,同时也可以向人类解释。确保所使用的BT在安全性和可靠性要求方面构造良好至关重要,特别是对于在关键环境中运行的机器人。在这项工作中,我们提出了一个正式规范的行为树和方法来证明已经使用的树的不变量,同时保持复杂的形式化的树简单的最终用户。允许测试行为树的特定实例,而不需要知道形式化的更抽象的层次。

英文摘要:

Behavior Trees (BT) are becoming increasingly popular in the robotics community. The BT tool is well suited for decision-making applications allowing a robot to perform complex behavior while being explainable to humans as well. Verifying that BTs used are well constructed with respect to safety and reliability requirements is essential, especially for robots operating in critical environments. In this work, we propose a formal specification of Behavior Trees and a methodology to prove invariants of already used trees, while keeping the complexity of the formalization of the tree simple for the final user. Allowing the possibility to test the particular instance of the behavior tree without the necessity to know the more abstract levels of the formalization.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑