Collaborative Research: An Audio-Based Spatiotemporal System for Automated Monitoring of Construction Operations

建筑行业一直缺乏实时性能监控、整体项目管理、劳动效率和预防浪费的工具。这导致近90%的建筑项目成本超支，平均比预测成本高出28%。预计该研究项目将促进建筑活动和运营跟踪和监测方式的变革。它将在未来的快速，非侵入性和成本效益的数据采集能力，由于使用音频信号，而不是主动传感器或数码相机的重要作用。最终，自动化项目监控系统所带来的效率、可持续性效益和降低的成本将使美国建筑、工程、施工和设施管理行业受益匪浅。该项目的成功也将对工程教育产生重大影响。该项目的教育活动与研究活动高度融合并相互关联。研究结果将用于制作教育材料，并将向其他机构的教育工作者公开提供。研究成果也将被纳入外联和参与活动，并将导致通过各种方案参与少数民族和代表性不足的群体。尽管最近在开发和实施音频信号处理技术来分析和建模复杂的系统和过程方面取得了进展，但音频信号的真实的附加值和潜在应用仍然不为土木工程研究界所知。该项目是首次尝试引入音频信号作为识别、跟踪和监控施工现场施工作业的替代信息源。目前用于识别和监控施工操作的方法要么是基于位置的，要么是基于机器视觉的，而隐含音频信号是克服其局限性的一个重大飞跃。该框架提供了通用信号处理和施工性能监测之间的缺失环节。该项目将扩大民用基础设施系统以及数字信号处理领域学者的研究视野。特别是，这一科学突破将为未来的研究奠定基础，自动识别和生命记录施工操作和设备动作，估计项目绩效指标，并制定纠正措施以保持项目绩效按计划进行。反过来，这将使未来开发新的自动化应用程序，用于施工顺序分析，生产力测量，项目监测和控制系统以及维护决策。

The construction industry has been suffering from the lack of real-time performance monitoring, holistic project management, labor efficiency, and waste-preventive tools. This has led to cost overruns in almost 90 percent of construction projects with an average of 28 percent higher than forecast costs. This research project is expected to facilitate a transformative change in the ways that construction activities and operations are tracked and monitored. It will play a significant role in future rapid, nonintrusive and cost effective data acquisition capacity due to the use of audio signals instead of active sensors or digital cameras. Eventually the incurred efficiencies, sustainability benefits, and reduced costs of an automated project monitoring system will significantly benefit the U.S. Architecture, Engineering, Construction and Facilities Management industry. Success in this project also promises significant impacts to engineering education. The project's educational activities are highly integrated and inter-related with the research activities. The research results will be used to create educational material and will be made publicly available to educators at other institutions. The research results will also be integrated into the outreach and engagement activities and will result in engaging minorities and underrepresented groups through various programs.Despite recent advances in developing and implementing audio signal processing techniques for analyzing and modeling complex systems and processes, the real added value and potential applications of audio signals are still unknown to the civil engineering research community. This project is the first attempt to introduce audio signals as an alternative source of information for recognizing, tracking and monitoring construction operations at jobsites. Current approaches for recognizing and monitoring construction operations are either location-based or machine-vision-based, and implying audio signals is a significant leap to overcome their limitations. The framework provides the missing link between generic signal processing and construction performance monitoring. This project will expand the research horizons for academics in civil infrastructure systems as well as in digital signal processing domains. Particularly, this scientific breakthrough will set the stage for future research in automatically identifying and life-logging construction operations and equipment actions, estimating project performance indices, and creating corrective measures to keep the project performance as planned. This, in turn, will enable the future development of novel, automated applications for construction sequence analysis, productivity measuring, project monitoring and control systems, and maintenance decision making.