DRU: Analysis of Current HAZMAT Response Capabilities and the Impact of Human Robot Interaction on Risk Mitigation

我们的研究解决了如何设计人机交互（HRI）来帮助决策者团队面对动态、不确定和危险的情况，这些情况需要快节奏的决策。我们目前的研究重点源于我们在人力资源研究领域的工作。我们一直在研究一个人和一个机器人如何协同工作，以解决安全关键应用中的问题，例如在机器人辅助城市搜索和救援（USAR）环境中；特别是，我们一直在研究HRI设计如何帮助人类决定在不久的将来采取什么行动，特别是在人类操作员看不到机器人的情况下。这种“远程”机器人操作本身就很困难，因为当机器人在环境中移动时，人类必须完全依赖于不断变化的参照系中呈现的信息。我们研究进展的下一个合乎逻辑的步骤是检查团队决策。为了研究多人/单机器人环境下的决策动力学，我们将研究危险物质（HAZMAT）处理。与USAR一样，HAZMAT通常要求机器人远离其操作人员。我们的工作将开发“在正确的时间提供正确的信息”的方法：这是决策者在压力下的长期目标。到目前为止，许多机器人界面都是由机器人专家设计的，为机器人专家设计的。随着机器人越来越商业化，人们开始更多地关注如何使界面对预期的最终用户可用。然而，据我们所知，以前没有人利用社会科学技术来确保满足机器人信息消费者的需求。这项研究汇集了来自机器人、人机交互、决策支持和社会科学界的研究人员。这种多学科的方法将使我们能够考虑人们如何相互作用，在快速变化的情况下需要的信息类型，以及我们如何评估和显示风险。在整个项目中，我们将直接与预期用户、第一响应者和决策者合作。现有的机器人系统将在演习中与第一响应者一起使用，向他们介绍最新的技术，提供快速采用的潜力。危险品领域的发现也将适用于USAR情况，如地震、飓风和恐怖袭击。

Our research addresses the question of how human-robot interaction (HRI) could be designed to assist teams of decision-makers in the face of dynamic, uncertain, and risky situations, where fast-paced decisions are required. Our current research focus grew out of our work in the HRI domain. We have been studying how one person and one robot work together to solve problems in safety-critical applications, such as in a robotic-assisted urban search and rescue (USAR) context; in particular, we have been examining how the HRI design can help the human decide what actions to take in the near future, especially in cases where the human operator cannot see the robot. Such "remote" robot operation is inherently difficult because the human must rely solely on information presented in a constantly changing frame of reference as the robot moves through the environment. The next logical step in our research progression is to examine team decision-making. To study decision-making dynamics in a multi-human/single-robot context, we will examine hazardous material (HAZMAT) handling. As with USAR, HAZMAT typically requires a robot to be deployed away from its human operator.Our work will develop methods for presenting the "right information at the right time": the perennial goal of the decision-maker under pressure. To date, many robotic interfaces have been designed by roboticists, for roboticists. With robots becoming more commercialized, more attention is starting to be paid to making interfaces usable by the intended end users. We know of no previous efforts, however, to use social science techniques to ensure that the needs of robot information consumers are met. This study brings together researchers from the robotics, human-computer interaction, decision support, and social science communities. This multidisciplinary approach will allow us to consider how people interact with one another, the types of information that are needed in rapidly changing situations, and how we can assess and display risk. We will work directly with the intended users, the first-responders and decision-makers, throughout the project. Existing robot systems will be used in exercises with first-responders to introduce them to the current state of the art, offering the potential for rapid adoption. Findings in the HAZMAT domain will also be applicable to USAR situations such as earthquakes, hurricanes, and terrorist attacks.