Development of a predictive safety control algorithm using laser scanners for excavators on construction sites

This paper presents a laser scanner–based predictive safety system for excavators. Blind spots on excavators and operators’ carelessness cause majority of fatal accidents, such as those in which deaths occur due to collisions with objects on a construction site. A proper safety system can enhance the safety of construction vehicles on construction sites. In this study, a safety control algorithm for collision avoidance was developed utilizing kinematics and a dynamic model based on the working area and object behavior predictions. The object behaviors were predicted by considering human pace states since excavators often operate in tandem with workers. Combining the data collected from static obstacles and moving objects, the researchers identified a safe working space. In the case of moving objects, the researchers predicted the probabilistic reachable area of workers via hypothesis testing and state estimation utilizing the estimated position and velocity information obtained by a laser scanner. Hypothesis testing was conducted to identify worker pace states, such as standing still, walking, jogging, and running, using estimated velocity. The working area was predicted via working part kinematic analysis of an excavator. Safety indices, such as time to collision (TTC) and warning index (x), were employed to define the safety level of an excavator in operation in the TTC–x domain. The safety level consists of safe, warning, and emergency braking levels. Furthermore, the researchers developed a control algorithm to avoid collision of the excavator with static and moving objects. Tests of the developed reachable area of a worker were conducted utilizing laser scanners. In addition, a simulation-based performance evaluation of the developed safety control algorithm was conducted with test results employing the excavator swing dynamic model.


  • English

Media Info

Subject/Index Terms

Filing Info

  • Accession Number: 01711715
  • Record Type: Publication
  • Files: TRIS
  • Created Date: Jun 19 2019 3:05PM