Data from: Toward AI-Augmented Engineering Workflows: Insights from the European Rover Challenge

Abstract The increasing adoption of Artificial Intelligence (AI) tools offers new opportunities to support engineering design and coordination, yet their most effective applications in complex workflows remain unclear. To investigate this, we collected survey data from teams participating in the European Rover Challenge (ERC) [1], an international robotics competition in which student teams design, build, and integrate complex planetary rover systems under strict technical, regulatory, and time constraints. This dataset comprises responses to a 40-question survey distributed to all 25 ERC 2025 teams, yielding 104 responses from 14 teams. The data capture team organization, development practices, AI usage, and challenges in rover design and integration. The analysis of these responses is presented by Sadik et al. [2]. Dataset organization The dataset consists of a single MS Excel file. Each row corresponds to the responses of an individual team member. Columns include anonymized team names, the role of the respondent within the team, and answers to the survey questions listed below. The survey was designed to capture perspectives from different roles within ERC teams using a role-dependent branching structure. Participants identified themselves as team leaders, group leaders, or team members and were directed to role-specific sets of questions. Team leaders responded to questions concerning team organization, experience, and knowledge transfer; group leaders focused on task assignment and coordination; and team members addressed daily execution and workflow-related practices. Following these role-specific sections, all respondents completed a shared block of questions covering core aspects of the development process, including task definition, dependencies, scheduling, integration activities, communication practices, and the use of AI tools. Although the survey comprised 40 core questions, several were instantiated in multiple contextual forms. Consequently, the dataset includes more columns, and the list below contains more items than the original number of survey questions. Response items Select the name of your team (2025): Select your role in the team: What was your team size (number of active members)? For how many years did your university participate in ERC/URC? How was the team organization structure? Are you using standard processes for your processes, which one and for which part? Do you plan ahead for knowledge transfer, and how do you make sure you gather all information needed for the future transfer? What share of the team 2024 was kept in the challenge 2025? Rate the knowledge transfer process: What are the biggest challenges in knowledge transfer? Which knowledge transfer methods were used? How do you usually assign tasks to your group members? What are the main problems you experienced in assigning tasks? How do you monitor whether tasks are progressing on time? How do you decide when a task is “completed”? What are the main problems you experienced in monitoring tasks? How was the team organization structure? How would you rank the effectiveness of the organization structure for your work? How were tasks assigned to you? Rate the level of quality of the following aspects: [Knowledge transfer process] Rate the level of quality of the following aspects: [Knowledge transfer content] Rate the level of quality of the following aspects: [Task definition] Rate the level of quality of the following aspects: [Technical constraints definition] Rate the level of quality of the following aspects: [Deadline and duration] Rate the level of quality of the following aspects: [Dependencies on other tasks] Rate the level of quality of the following aspects: [Acceptance/done criteria] Rate the level of quality of the following aspects: [Time to design] Rate the level of quality of the following aspects: [Time to manufacture] Rate the level of quality of the following aspects: [Time to perform integration] Rate the level of quality of the following aspects: [Time to perform test] For how many years do you personally work at ADC? Where are you involved? Estimate time spent in different phases: [Recruitment & Onboarding] Estimate time spent in different phases: [Knowledge Transfer] Estimate time spent in different phases: [Background Research] Estimate time spent in different phases: [Requirements & Concept (Pre-selection)] Estimate time spent in different phases: [Requirement & Concept (Post-selection)] Estimate time spent in different phases: [Detailed Design] Estimate time spent in different phases: [Development] Estimate time spent in different phases: [Testing & Validation] Estimate time spent in different phases: [Final Testing before Final Event] Estimate time spent in different phases: [Documentation] Estimate time spent in different phases: [Sponsor Search (Marketing)] Estimate time spent in different phases: [Public Relation] A posteriori, do you think any of these tasks could be made more efficiently? [Recruitment & Onboarding] A posteriori, do you think any of these tasks could be made more efficiently? [Knowledge Transfer] A posteriori, do you think any of these tasks could be made more efficiently? [Background Research] A posteriori, do you think any of these tasks could be made more efficiently? [Requirements & Concept (Pre-selection)] A posteriori, do you think any of these tasks could have been made more efficiently? [Requirement & Concept (Post-selection)] A posteriori, do you think any of these tasks could be made more efficiently? [Detailed Design] A posteriori, do you think any of these tasks could be made more efficiently? [Development] A posteriori, do you think any of these tasks could be made more efficiently? [Testing & Validation] A posteriori, do you think any of these tasks could have been made more efficiently? [Final Testing before Final Event] A posteriori, do you think any of these tasks could be made more efficiently? [Documentation] A posteriori, do you think any of these tasks could have been made more efficiently? [Sponsor Search (Marketing)] A posteriori, do you think any of these tasks could have been made more efficiently? [Public Relation] Percentage of work requiring rework/revision in term of time: [Concept & Requirements] Percentage of work requiring rework/revision in term of time: [Design] Percentage of work requiring rework/revision in term of time: [Development] Percentage of work requiring rework/revision in term of time: [Integration] Percentage of work requiring rework/revision in term of time: [Testing] Percentage of work requiring rework/revision in term of time: [Documentation] Which were the most recurring loops that result in wasted effort? What caused the biggest delays? How do you track requirements and with which tools? Do you use AI Tools (ChatGPT, Claude,...) in your development process? Which AI tool and for which purpose? How is communication organized? How often does communication take place? Which tools did you use to communicate/coordinate? Which communication problems affected you most often? What was the biggest cause of conflicts? What was the most impacting issue that affected the quality of the final product? What would you change in your team’s strategy? If an AI assistant could solve one organizational problem for you, what should it do? If an AI assistant could solve one technical problem for you, what should it do? Data anonymization statement This dataset contains anonymized survey responses from student engineering teams. No personal identifiers (names, emails, institutions) are included. Free-text responses were reviewed and generalized to prevent indirect identification (e.g., geographic references). Acknowledgements The authors gratefully acknowledge the organizers of the ERC, especially the European Space Foundation under the leadership of Łukasz Wilczyński, for their support and for facilitating access to the participating teams. We also wish to express our sincere thanks to the members, group leads, and team leads of the ERC 2025 teams who generously shared their time and perspectives by completing the survey: 4Space (Spain), AAU Space Robotics (Denmark), AGH Space Systems (Poland), ASU ROAR (Egypt), DJS Antariksh (India), FHNW Rover Team (Switzerland), FRoST (Germany), KNR Rover Team (Poland), Mars Rover Manipal (India), Orion Team (Poland), OzU Rover Team (Turkey), ProjectRED (Italy), Sapienza Technology Team (Italy), UPC Space Program (Spain), WARRYZN Space Robotics (Germany). This work would not have been possible without their openness in sharing their experiences and challenges. References [1] European Rover Challenge (ERC) – official website. https://roverchallenge.eu/, accessed: 2025-02-19 [2] Sadik, A.R., Joublin, F., Bujny, M., Ceravola, A., Smith, J.: Toward AI-augmented Engineering Workflows: Insights from the European Rover Challenge.摘要 人工智能（Artificial Intelligence）工具的日益普及为工程设计与协同工作带来了新的机遇，但其在复杂工作流中的最优应用场景仍不明晰。为探究这一问题，我们从参与欧洲火星车挑战赛（European Rover Challenge，ERC）[1]的团队处收集了调研数据。该赛事为国际性机器人竞赛，要求学生团队在严苛的技术、法规与时间约束下，设计、建造并集成复杂的行星探测车系统。本数据集包含面向ERC 2025全部25支参赛团队发放的40题调查问卷的回复，最终回收了来自14支团队的104份有效问卷。数据涵盖团队组织架构、开发实践、人工智能工具使用情况，以及火星车设计与集成过程中面临的挑战。相关回复的分析结果已由Sadik等人[2]发表。 数据集组织形式 本数据集仅包含一个Microsoft Excel（MS Excel）文件。每一行对应一位团队成员的调研回复，列字段包括匿名化后的团队名称、受访者在团队中的角色，以及下述调查问卷问题的答案。 本次调研采用基于角色的分支式问卷结构，旨在收集ERC参赛团队不同岗位成员的视角。受访者需先标注自身身份为团队负责人、小组负责人或普通团队成员，随后跳转至对应角色的专属问题模块：团队负责人需回答涉及团队组织架构、项目经验与知识转移（Knowledge Transfer）的问题；小组负责人需聚焦任务分配与协同相关内容；普通团队成员则需汇报日常执行与工作流相关的实践情况。在完成角色专属模块后，所有受访者均需完成一套通用问题模块，涵盖开发流程的核心环节，包括任务定义、依赖关系、进度安排、集成活动、沟通实践以及人工智能工具的使用情况。 尽管调研仅包含40道核心问题，但部分问题会根据场景衍生出多个变体。因此，数据集的列字段数量多于原始问卷的问题总数，下述响应条目列表的条目数也多于原始问卷的问题量。 响应条目 选择您所属的2025赛季参赛团队名称： 选择您在团队中的角色： 您的团队规模（活跃成员人数）为多少？ 您的大学参与ERC/URC赛事已有多少年？ 团队的组织架构形式如何？ 您是否为团队流程采用了标准规范？若有，具体采用了哪些规范，应用于哪些环节？ 您是否为知识转移（Knowledge Transfer）制定了前置规划？如何确保收集到未来知识转移所需的全部信息？ 2024赛季团队中有多少比例的成员保留至2025赛季赛事？ 请对知识转移流程进行评分： 知识转移过程中面临的最大挑战是什么？ 采用了哪些知识转移方法？ 您通常如何为小组成员分配任务？ 任务分配过程中遇到的主要问题有哪些？ 您如何监控任务是否按计划进度推进？ 您如何判定一项任务已“完成”？ 任务监控过程中遇到的主要问题有哪些？ 团队的组织架构形式如何？ 您如何评价当前组织架构对自身工作的有效性？ 任务是如何分配给您的？ 请对下述各方面的质量水平进行评分：[知识转移流程] 请对下述各方面的质量水平进行评分：[知识转移内容] 请对下述各方面的质量水平进行评分：[任务定义] 请对下述各方面的质量水平进行评分：[技术约束定义] 请对下述各方面的质量水平进行评分：[截止日期与工期] 请对下述各方面的质量水平进行评分：[任务间依赖关系] 请对下述各方面的质量水平进行评分：[验收/完成标准] 请对下述各方面的质量水平进行评分：[设计阶段耗时] 请对下述各方面的质量水平进行评分：[制造阶段耗时] 请对下述各方面的质量水平进行评分：[集成活动耗时] 请对下述各方面的质量水平进行评分：[测试阶段耗时] 您个人在ADC领域的工作年限为多少年？ 您参与的工作内容是什么？ 估算各阶段花费的时间：[招聘与入职培训（Recruitment & Onboarding）] 估算各阶段花费的时间：[知识转移（Knowledge Transfer）] 估算各阶段花费的时间：[背景调研（Background Research）] 估算各阶段花费的时间：[需求与概念设计（预选阶段）（Requirements & Concept (Pre-selection)）] 估算各阶段花费的时间：[需求与概念设计（选后阶段）（Requirement & Concept (Post-selection)）] 估算各阶段花费的时间：[详细设计（Detailed Design）] 估算各阶段花费的时间：[开发（Development）] 估算各阶段花费的时间：[测试与验证（Testing & Validation）] 估算各阶段花费的时间：[最终赛事前的最终测试（Final Testing before Final Event）] 估算各阶段花费的时间：[文档编制（Documentation）] 估算各阶段花费的时间：[赞助商招募（营销）（Sponsor Search (Marketing)）] 估算各阶段花费的时间：[公共关系（Public Relation）] 事后复盘来看，您认为下列哪些任务可以更高效地完成？[招聘与入职培训] 事后复盘来看，您认为下列哪些任务可以更高效地完成？[知识转移] 事后复盘来看，您认为下列哪些任务可以更高效地完成？[背景调研] 事后复盘来看，您认为下列哪些任务可以更高效地完成？[需求与概念设计（预选阶段）] 事后复盘来看，您认为下列哪些任务可以更高效地完成？[需求与概念设计（选后阶段）] 事后复盘来看，您认为下列哪些任务可以更高效地完成？[详细设计] 事后复盘来看，您认为下列哪些任务可以更高效地完成？[开发] 事后复盘来看，您认为下列哪些任务可以更高效地完成？[测试与验证] 事后复盘来看，您认为下列哪些任务可以更高效地完成？[最终赛事前的最终测试] 事后复盘来看，您认为下列哪些任务可以更高效地完成？[文档编制] 事后复盘来看，您认为下列哪些任务可以更高效地完成？[赞助商招募（营销）] 事后复盘来看，您认为下列哪些任务可以更高效地完成？[公共关系] 按时间占比统计，需要返工/修订的工作比例：[概念设计与需求阶段] 按时间占比统计，需要返工/修订的工作比例：[设计阶段] 按时间占比统计，需要返工/修订的工作比例：[开发阶段] 按时间占比统计，需要返工/修订的工作比例：[集成阶段] 按时间占比统计，需要返工/修订的工作比例：[测试阶段] 按时间占比统计，需要返工/修订的工作比例：[文档编制阶段] 导致大量无效工作量的最常见循环环节是什么？ 造成最大进度延误的原因是什么？ 您如何跟踪需求？使用了哪些工具？ 您是否在开发过程中使用人工智能工具（如ChatGPT、Claude等）？ 使用了哪些人工智能工具？具体用于哪些场景？ 团队的沟通组织形式是怎样的？ 沟通的频率如何？ 您使用了哪些工具进行沟通与协同？ 最常遇到的沟通问题是什么？ 引发冲突的最主要原因是什么？ 对最终产品质量影响最大的问题是什么？ 您会对团队的战略做出哪些调整？ 如果一款AI智能体（AI Agent）可以帮您解决一项组织管理问题，它应当实现什么功能？ 如果一款AI智能体可以帮您解决一项技术问题，它应当实现什么功能？ 数据匿名化声明 本数据集包含学生工程团队的匿名化调研回复，未包含任何个人标识信息（如姓名、邮箱、所属机构）。针对自由文本回复，我们已进行审查与泛化处理，以避免间接识别风险（例如通过地理位置信息）。 致谢 作者谨向欧洲火星车挑战赛的主办方，尤其是在Łukasz Wilczyński领导下的欧洲空间基金会（European Space Foundation）致谢，感谢其提供的支持与协助，使我们得以接触参赛团队。同时，我们衷心感谢ERC 2025所有参赛团队的成员、小组负责人与团队负责人，感谢他们抽出宝贵时间完成问卷并分享自身经验与挑战，这些团队包括：4Space（西班牙）、AAU Space Robotics（丹麦）、AGH Space Systems（波兰）、ASU ROAR（埃及）、DJS Antariksh（印度）、FHNW Rover Team（瑞士）、FRoST（德国）、KNR Rover Team（波兰）、Mars Rover Manipal（印度）、Orion Team（波兰）、OzU Rover Team（土耳其）、ProjectRED（意大利）、Sapienza Technology Team（意大利）、UPC Space Program（西班牙）、WARRYZN Space Robotics（德国）。若无他们坦诚分享自身经历与挑战，本研究无法顺利完成。 参考文献 [1] 欧洲火星车挑战赛（European Rover Challenge，ERC）官方网站：https://roverchallenge.eu/，访问时间：2025-02-19 [2] Sadik, A.R.、Joublin, F.、Bujny, M.、Ceravola, A.、Smith, J.：面向人工智能增强型工程工作流：来自欧洲火星车挑战赛的启示。