Real-Time Guidance for Large-Scale Construction and Assembly Tasks

点击翻译按钮获取中文摘要

Imagine that you have volunteered to help in the construction of a house for a non-profit company that builds affordable homes for low-income families. You arrive on the construction site eager to help, but quickly find that the instructions provided to you by the site manager go beyond your knowledge, and the tools you are asked to use are unfamiliar. As you struggle through the day your frustration builds, and you are constantly asking for guidance and assistance from others, distracting them from their tasks as well. You can't help but feel like your volunteer experience has been more of a hindrance than a help. But what if, when you had arrived, you had received a smartwatch, that recognized your skill level, automatically assigned you to appropriate tasks, and provided step-by-step instructions while you worked. This example demonstrates the learning and guidance challenges for one specific but important scenario. However, the overall challenge of supporting learning and guidance within coordinated assembly efforts applies to a broad range of domains, such as habitat construction, home improvement, furniture assembly, disaster relief, and refugee shelter construction. We focus on how to enable intelligent construction and assembly environments, where computational technologies, such as sensors and mobile devices, are used to provide guidance and coordination in real-time, and adapt to the worker's needs and environmental changes. To realize this future it will be critical to develop methods to recognize a worker's skill level, technologies to monitor their progress on a task, and interfaces to deliver instructions in real time on mobile and wearable devices. Research in computer-supported collaborative work has focused on enabling efficient coordination for remote collaboration in the virtual realm. With this work, we aim to extend these concepts into the physical realm, supporting collocated coordination. Furthermore, research in software learning has looked at providing effective tutorials for assisting with procedural tasks when using desktop software. Our work will advance such findings to physical construction and assembly tasks. We are developing interaction methods to allow novice and expert workers to coordinate efficiently to accomplish their tasks, using multiple forms of sensors and mobile and wearable devices. In addition, we are developing systems and interfaces to allow project managers to monitor the progress of a build, and intelligently distribute subtasks to workers based on their individual skills. We are also investigating low cost-technologies which can provide these affordances in developing regions, where efficient habitat construction is critical. By augmenting such construction and assembly environments with digital technologies and interfaces, people of all skills and backgrounds will be empowered to participate, learn and succeed.