Current methods for spatial information acquisition and modeling rely on expensive laser range scanners that produce dense point clouds which require hours or days of post-processing to arrive at a finished model. While these methods produce very detailed 3D models of the scanned scene, the associated computational time burden precludes these methods from being used onsite for real-time decision-making. A rapid modeling approach for modeling local scenes for construction equipment operations, developed by the investigators with previous NSF support, presents significant advantages over full range scanning that require computationally intensive image processing. This approach provided a foundation for the development of the proposed novel methods of acquiring, integrating, modeling, and analyzing project site spatial data, including dynamic site information (motion of personnel and equipment), that allow scalability and robustness for real-time field deployment. These methods will enable complete local area modeling in the order of seconds, and with sufficient accuracy for applications such as real-time safety monitoring and advanced equipment control. The development of a real-time obstacle avoidance system using rapidly generated site models is also proposed, as well as an experimental plan with real infrastructure operation sites and equipment, to validate the proposed site data acquisition and obstacle avoidance methods. It is expected that the proposed approach will result in significant equipment safety improvements while at the same time lessening the need for skilled workers to operate heavy equipment in a wide range of site working conditions. Initial experimental work has demonstrated the feasibility of this approach.
The construction industry accounts for a large percentage of workforce fatalities in the United States. An Occupational Safety and Health Administration (OSHA) study of on-the-job fatalities in the construction industry in the U.S. has shown that over 50% of the accidents stemmed largely from operation of heavy equipment. One of the major causes of these accidents is the lack of safety features installed on heavy equipment that is currently in use. Great potential exists in making heavy equipment safer and more efficient by developing a system that would integrate effective acquisition and handling of externally sensed site spatial information in real-time. With the internal sensors and computing power available commercially now to retrofit equipment, such systems are technically and economically feasible. The aim of the proposed research is therefore to develop methods and technologies that would allow an operator to operate equipment more safely and at higher speeds in cluttered environments, even in situations where visibility is poor as in underground work. The technologies proposed by the research team would assist in minimizing the impact of key types of human error and will have the potential to radically improve safety and efficiency of infrastructure operations that involve heavy equipment.
