Evaluating the Impact of Location-Aware Sensor Data Imperfections on Autonomous Jobsite Safety Monitoring

Abstract

Construction is one of the industries with the highest fatality, non-fatal injuries and illness rates across the United States (CPWR, 2008). Although construction professionals have made great efforts to improve safety training, site supervision and design for safety (among others), the safety performance in construction is still not satisfactory due to human errors and lack of situational awareness during construction activities. Recent advances in sensing and computing technologies offer a solution for improving safety performance by providing rich information about location and, hence, worker safety. Our research envisions an autonomous jobsite safety monitoring system which utilizes distributed data and information collected from various sources (RFID, localization sensors, accelerometers, load cell sensors and building information models). The research is divided into three phases: (1) knowledge elicitation; (2) development and deployment of the autonomous safety monitoring system in a distributed computing environment; and (3) exploring the impacts of data imperfections situations (e.g., erroneous or missing data) on the safety monitoring system and approaches to reduce the impacts of the imperfections on the safety system. Prior research has reported on the first two phases (Julien, et al, 2005; Luo, et al, 2011; O'Brien, et al, 2008). This paper reports the preliminary results of evaluating the impacts of data imperfections on autonomous safety monitoring using data from location-aware sensors, which provides the most important information for the autonomous jobsite safety monitoring system. The rest of this paper is organized as follows: The Related Work section summarizes the location-aware jobsite safety monitoring applications in construction and the imperfect data collected from sensors. The Research Approach section describes the research approach, including the testbed used for data collection, the test case used to evaluate the impacts of imperfect data on autonomous safety monitoring and the process of how the data is used to evaluate the impacts. The Preliminary Results section reports the preliminary results of the impacts on autonomous safety monitoring. Finally, the Conclusion section summarizes the results and proposes future research suggestions.