Toward the Next Generation Jobsite Safety: A More Accurate Autonomous Monitoring
DescriptionBy 2025, global construction is expected to grow 70% according to Global Construction andOxford Economics. With this growth comes a need to focus on jobsites safety and efficiencybecause the construction industry has one of the highest accident and fatality rates among othermajor industries, with more than 60,000 fatal accidents each year worldwide. In Hong Kong,construction job site injuries account for nearly 20% of all work related fatalities and injuries. Asone of the most dangerous jobs in the world, this needs should be addressed swiftly.In order to deal with the increasingly complex and challenging construction environment, recentstudies have focused on adopting sensing and communication technologies to revolutionize theway of traditional construction safety management. Such applications can provide workers withthe communication and information tools for effectively identifying and reporting safety hazardson jobsites. This is a proactive approach that helps workers improve situational awarenessthrough an early warning mechanism, for example when a construction worker is exposed todangerous hazards, alarms will be triggered automatically for both the worker and the supervisor.However, these applications were developed under the assumption that the data collected fromsensors is perfect, which is not true. Although sensor networks can provide many benefits, theyare more susceptible to malfunctions that can result in imperfect data. Imperfect data may lead tomissing detections or false alarms, which drastically reduces the significance of the use ofsensors and even results in severe consequences Therefore; there is a great need to create newsolutions to make the existing ones more robust, accurate and precise.The research objective of this proposal is to seek new approaches to develop a more reliable andhigh- performance system for construction safety management. Particular attention focuses onthe data processing and decision support level. Data fusion, stochastic programming andoptimization techniques will be adopted for algorithm and decision making mechanismdevelopment in this project. A testbed with sensing devices to represent the intelligentconstruction jobsite will be designed and implemented to test and validate the performance ofour proposed improvement mechanisms for jobsite autonomous safety monitoring systems.In terms of broader impact, our application can significantly benefit a large industry worldwidefrom improvements in practices. In addition, this research has considerable potential to impactother domains (e.g. productivity analysis) since the problems of imperfect data are generic ininformation extraction and decision making, Moreover, to ensure significant impact, we willcombine the project’s research efforts with both educational efforts and industrial programs.
|Effective start/end date||1/01/16 → 24/12/19|