The construction industry is infamously recognized for having jobsites identified as hazardous working environments due to workers-on-foot commonly functioning at dangerous heights and in close proximity to other construction entities, among other reasons. These hazardous operations result in an increased risk of worker injuries and fatalities caused by fall hazards and struck-by object or equipment incidents. Construction workers are involved in these incidents due to proximity and visibility issues caused by the congested, continuously changing, and dynamic nature of the worksite. Previous research studies concentrated on developing various resource localization systems to prevent these hazardous interactions on site based on proximity of workers to hazards on site. However, these previously established systems typically generate false positive alarms that produce the false-alarm effect for workers and contributes to the risks present on jobsites. This thesis develops a system to minimize the occurrence of false proximity alarms on construction worksites for proximity and visibility related hazards.
Information was collected from contractor interactions and literature reviews to create a safety rules-based system to monitor the interactions on site pertaining to falls and struck-by equipment incidents. Along with the developed rules-based system, an enhanced proximity detection system that uses real-time sensor data was devised to track worker-on-foot’s field-of-view and proximity to other construction entities. A virtual model was also added to the study to enable spatial analysis of the system and entities in the real-world. This model contributed to monitoring and observing hazardous construction interactions without liability issues brought forth by real-world experiments. Ultimately, the developed framework issued safety warnings to workers when they had their backs facing the other entities on site, which created restricted visibility conditions for hazards in their close proximity. This focus reduced the emission of false positive alarms while still successfully alerting workers of fall and struck-by hazards. The developed detection system and virtual model has great potential to enhance construction safety by monitoring field-of-view to reduce false positive alarms and mitigate the occurrence of visibility and proximity related incidents on worksites.