NRI: INT: COLLAB: Program Verification and Synthesis for Collaborative Robots

NRI：INT：COLLAB：协作机器人的程序验证和综合

• 批准号: 1924435
• 负责人: Maya Cakmak
• 金额: $ 54.09万
• 依托单位: University of Washington
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2019
• 资助国家: 美国
• 起止时间: 2019-09-01 至 2023-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1924435&HistoricalAwards=false
• 关键词: NRI; INT; COLLAB; Program; Verification

Developing applications for robots currently requires advanced knowledge and skills in both robotics and software engineering. As a result, development of robot applications is currently only accessible to a very small segment of the population, excluding those who might have the deepest insight about potential application domains for robots (e.g., experts in logistics, manufacturing, marketing), those who might be better suited to design robot behaviors for human interaction (e.g., designers, artists, psychologists), and those who might be using and repurposing robots day-to-day (e.g., end-users, supervisors, maintenance personnel). This project aims to design and implement a suite of tools and methods, powered by developments in the field of programming languages, to enable the programming of interactive robots with minimal prior experience. These tools and methods will assist all stages of the programming process, including ideation of what the robot behavior should be, coding of the program, its testing on the robot to guarantee certain properties, and adapting the program for deployments in different contexts. The research will make robot programming accessible to the broader public and "democratize" it by substantially lowering barriers to engagement in technology development. Through collaborations with domain experts, the principles, methods, and tools created by the project will be applied to real-world robotics applications, transferring the scientific contributions of the project to robotics organizations. Through an integrated education program, computer science training and curriculum will be enhanced across three institutions by introducing new courses and course modules on robotics software development. To facilitate public understanding of and engagement with science and technology, a robot programming environment for elementary- and middle-school-aged children will be developed along with a structured curriculum for its use in an outreach summer program, and these will be widely disseminated as open-source materials.The key innovation of the project is the development of novel paradigms for programming that use state-of-the-art program analysis and synthesis methods to support a wide range of programming tasks across the development cycle for robotics applications. Specifically, the project will support each stage of the cycle in three research thrusts with the following objectives: (1) facilitating the capture of user intent by translating designer and end-user intent captured through acting and dialogue into robot programs; (2) supporting iterative development and testing by facilitating the development, refinement, analysis, and synthesis of robust, reliable, and effective robot programs; (3) optimizing deployment by enabling the adaptation of existing robot programs into new environments, scenarios of use, and social contexts. The outcomes of these research thrusts will be integrated into a suite of robot programming tools within the ROS framework and deployed in real-world robot programming tasks in collaboration with industrial partners.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.

目前，开发机器人应用程序需要机器人技术和软件工程方面的先进知识和技能。因此，机器人应用程序的开发目前只对很小一部分人开放，不包括那些可能对机器人潜在应用领域有最深见解的人（例如，物流，制造，营销专家），那些可能更适合为人类交互设计机器人行为的人（例如，设计师，艺术家，心理学家），以及那些可能每天都在使用和重新利用机器人的人（例如，最终用户，主管，管理人员，管理人员）。维修人员)。该项目旨在设计和实现一套工具和方法，由编程语言领域的发展提供动力，使交互式机器人的编程具有最小的先验经验。这些工具和方法将协助编程过程的所有阶段，包括机器人行为的构思，程序的编码，在机器人上进行测试以保证某些属性，以及调整程序以适应不同环境的部署。这项研究将使机器人编程对更广泛的公众开放，并通过大幅降低参与技术开发的障碍使其“民主化”。通过与领域专家的合作，项目创建的原理、方法和工具将应用于现实世界的机器人应用，将项目的科学贡献转移到机器人组织。通过综合教育计划，计算机科学培训和课程将通过引入机器人软件开发的新课程和课程模块，在三所大学之间得到加强。为了促进公众对科学技术的理解和参与，将为小学和中学儿童开发一个机器人编程环境，并为其在夏季推广项目中使用的结构化课程，这些将作为开源材料广泛传播。该项目的关键创新是开发新的编程范例，使用最先进的程序分析和综合方法来支持机器人应用程序开发周期中的广泛编程任务。具体而言，该项目将在三个研究重点中支持周期的每个阶段，目标如下：(1)通过将通过表演和对话捕获的设计师和最终用户意图翻译成机器人程序，从而促进用户意图的捕获；(2)通过促进健壮、可靠和有效的机器人程序的开发、改进、分析和综合，支持迭代开发和测试；(3)优化部署，使现有机器人程序适应新的环境、使用场景和社会背景。这些研究成果将集成到ROS框架内的一套机器人编程工具中，并与工业伙伴合作部署在现实世界的机器人编程任务中。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

项目成果

Following Natural Language Instructions for Household Tasks With Landmark Guided Search and Reinforced Pose Adjustment

• DOI: 10.1109/lra.2022.3178804
• 发表时间: 2022-07
• 期刊: IEEE Robotics and Automation Letters
• 影响因子: 5.2
• 作者: Michael Murray;M. Cakmak

Interaction Templates: A Data-Driven Approach for Authoring Robot Programs

交互模板：用于编写机器人程序的数据驱动方法

• DOI: 10.1184/r1/20031965.v1
• 发表时间: 2022
• 期刊: PLATEAU: 12th Annual Workshop at theIntersection of PL and HCI
• 作者: Porfirio David;Cakmak, Maya;Sauppé, Allison;Albarghouthi, Aws;Mutlu, Bilge
• 通讯作者: Mutlu, Bilge

An Exploration of Accessible Remote Tele-operation for Assistive Mobile Manipulators in the Home

• DOI: 10.1109/ro-man50785.2021.9515511
• 发表时间: 2021-08
• 期刊: 2021 30th IEEE International Conference on Robot & Human Interactive Communication (RO-MAN)
• 作者: M. E. Cabrera;T. Bhattacharjee;Kavi Dey;M. Cakmak

Mimic: In-Situ Recording and Re-Use of Demonstrations to Support Robot Teleoperation

模仿：现场记录和重复使用演示以支持机器人远程操作

• DOI: 10.1145/3526113.3545639
• 发表时间: 2022
• 期刊: ACM Symposium on User Interface Software and Technology (UIST
• 作者: Mahadevan, Karthik;Chen, Yan;Cakmak, Maya;Tang, Anthony;Grossman, Tovi
• 通讯作者: Grossman, Tovi

Influencing Behavioral Attributions to Robot Motion During Task Execution

在任务执行过程中影响机器人运动的行为归因

• 发表时间: 2021
• 期刊: Conference on Robot Learning (CoRL
• 作者: Nick Walker, Christoforos Mavrogiannis