LAUSR

Monitoring and analyzing the operations of construction equipment is critical in the construction engineering and management domain. The ability to detect and classify major activities that construction equipment performs can support a project manager in making proper project-related decisions such as resource allocation and scheduling, resulting in improved productivity. Earth-moving activities as performed by an excavator, which is one of the most frequently used pieces of construction equipment, are normally repetitive by nature and possess unique features in terms of their patterns of operation. This study develops an activity identification algorithm capable of automatically classifying predefined earth-moving activities that are currently in progress. Given that the excavator is operated using joysticks, the joystick signals include unique patterns that exhibit the similar overall shapes but may not uniformly line up with time for a specific activity. The proposed study examines a dynamic time warping algorithm that determines similarities between a predefined activity and a measured signal distorted in time. The feasibility of the algorithm is verified through experiments involving activities such as digging, leveling, lifting, and trenching that were easily and accurately identified by the algorithm. The proposed task-identification algorithm could be used to develop an automated system of establishing machine parameters and to calculate the durations of operations and cycle times.