Designing planning and control interfaces to support user collaboration with flying robots
Designing planning and control interfaces to support user collaboration with flying robots
复制标题
设计规划和控制界面以支持用户与飞行机器人的协作
DOI：
--
复制
发表时间：
2017
期刊：
影响因子：
--
通讯作者：
T. Fong
中科院分区：
文献类型：
--
作者：
D. Szafir;Bilge Mutlu;T. Fong
机器人正变得越来越普遍，并且已经在各种活动中提供帮助，从太空探索到家务劳动。传感器、马达和微机电系统设计方面的最新进展使新型小型空中机器人得以发展。这些自由飞行的机器人作为移动传感器平台，在难以进入或无法测量的地区收集数据，在帮助人类方面具有很大的前景。在这项工作中，我们探索了界面的设计，支持用户与自由飞行的机器人一起完成任务，包括库存物流和管理，环境数据收集和视觉检查。扩展先前在地面机器人控制接口方面的工作，我们进行了一项形成性研究，以确定自由飞行接口的关键设计要求。我们设计了几个现实的任务，用于评估室内自由飞行操作环境下的人机交互。我们实现了三个原型接口，每个接口都提供了不同程度的支持，使远程用户能够与飞行机器人一起工作，以规划、交流目标、完成任务和响应动态环境中的变化。对每个界面的实验评估发现，与支持低级远程操作或基于路径点的监督控制的界面相比，设计用于支持使用交互式时间轴和三维空间路径点的协作规划和重新规划的界面显着提高了用户完成任务的效率，他们对任务需求自发变化的干预能力，以及他们对机器人作为队友的评价。我们的研究结果证明了数据驱动设计过程的实用性，并表明除了任务执行之外，自由飞行界面还需要考虑规划阶段。此外，我们还展示了在出现计划外事件时为中断机器人操作提供接口支持的重要性。
Robots are becoming increasingly prevalent and are already providing assistance in a variety of activities, ranging from space exploration to domestic housework. Recent advances in the design of sensors, motors, and microelectromechanical systems have enabled the development of a new class of small aerial robots. These free-flying robots hold great promise in assisting humans by acting as mobile sensor platforms to collect data in areas that are difficult to access or infeasible to instrument. In this work, we explored the design of interfaces that support users in working with free-flying robots to accomplish tasks including inventory logistics and management, environmental data collection, and visual inspection. Extending prior work in control interfaces for ground robots, we conducted a formative study in order to identify key design requirements for free-flyer interfaces. We designed several realistic tasks for use in evaluating human–robot interaction within the context of indoor free-flyer operation. We implemented three prototype interfaces that each provide varying degrees of support in enabling remote users to work with a flying robot to plan, communicate goals, accomplish tasks, and respond to changes in a dynamic environment. An experimental evaluation of each interface found that the interface designed to support collaborative planning and replanning using an interactive timeline and three-dimensional spatial waypoints significantly improved users’ efficiency in accomplishing tasks, their ability to intervene in response to spontaneous changes in task demands, and their ratings of the robot as a teammate compared to interfaces that support low-level teleoperation or waypoint-based supervisory control. Our results demonstrate the utility of a data-driven design process and show the need for free-flyer interfaces to consider planning phases in addition to task execution. In addition, we demonstrate the importance of providing interface support for interrupting robot operations as unplanned events arise.