Collaborative Research: SHF: Small: Interactive Synthesis and Repair For Robot Programs

合作研究：SHF：小型：机器人程序的交互式合成和修复

• 批准号: 2102291
• 负责人: Arjun Guha
• 金额: $ 19.38万
• 依托单位: Northeastern University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2020
• 资助国家: 美国
• 起止时间: 2020-09-01 至 2024-05-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2102291&HistoricalAwards=false
• 关键词: Collaborative; Research; SHF; Small; Interactive

Over the past few years, robots have started to be deployed in unstructured human environments. There are hundreds of robots deployed in hospitals, hotels, and supermarkets. Unfortunately, the software that runs on robots is programmed using low-level abstractions and languages, and is hard to transfer across robots and environments. In addition robotic software requires complex control logic to ensure that robots are safe and well-behaved in all situations. Thus, robot software is extraordinarily hard to write and maintain. This research project develops tools and techniques to make robot software safer, easier to write, and easier to maintain. The intellectual merits of the project are the development of (1) techniques for fixing bugs in robot software, based on advances to automatic program repair and program synthesis; (2) abstractions for writing robot software that can automatically handle certain kinds of failures, based on new programming-language design; (3) methods for checking the correctness of robot software, based on new program-verification techniques. The project's broader significance and importance are that it helps make robot software easier to write and maintain, and cheaper, safer, and more reliable. The project encourages further research at the intersection of programming languages and robotics by publishing research results and releasing open-source software. The project also involves high-school outreach workshops to broaden participation in computing.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

在过去的几年中，机器人开始部署在非结构化的人类环境中。在医院，酒店和超市中部署了数百个机器人。不幸的是，在机器人上运行的软件是使用低级抽象和语言编程的，并且很难在机器人和环境中传输。此外，机器人软件还需要复杂的控制逻辑，以确保机器人在所有情况下都安全且行为良好。因此，机器人软件非常难以编写和维护。该研究项目开发了工具和技术，以使机器人软件更安全，更易于编写，并且更易于维护。该项目的智力优点是基于自动程序维修和程序合成的进步，开发了（1）用于修复机器人软件中错误的技术； （2）编写机器人软件的抽象可以根据新的编程设计设计自动处理某些类型的失败； （3）基于新的程序验证技术检查机器人软件正确性的方法。该项目更广泛的意义和重要性是，它有助于使机器人软件更易于编写和维护，更便宜，更安全，更可靠。该项目通过发布研究结果并释放开源软件来鼓励在编程语言和机器人技术的交集上进行进一步的研究。该项目还涉及高中宣传研讨会以扩大计算的参与。该奖项反映了NSF的法定任务，并且使用基金会的知识分子优点和更广泛的影响审查标准，被认为值得通过评估来获得支持。

项目成果

Robot Action Selection Learning via Layered Dimension Informed Program Synthesis

• 发表时间: 2020-08
• 作者: Jarrett Holtz;Arjun Guha;Joydeep Biswas

Do Machine Learning Models Produce TypeScript Types that Type Check?

机器学习模型是否会生成可进行类型检查的 TypeScript 类型？

• 发表时间: 2023
• 期刊: European Conference on Object Oriented Programming (ECOOP
• 作者: Ming-Ho Yee;Arjun Guha
• 通讯作者: Arjun Guha