Tripping risk evaluation system based on human behavior simulation in laser-scanned 3D as-is environments

Abstract

Abstract Tripping-related falls are the leading cause of unintentional deaths in the elderly. Tripping risk must be evaluated by considering intrinsic and extrinsic factors affecting the interactions between humans and the environment. The goal of this study is to develop a new tripping risk evaluation system based on human behavior simulation in an “as-is” environment model, i.e., environment model reflecting detailed environmental geometry. First, a three-dimensional (3D) as-is environment model is automatically constructed from laser-scanned point clouds in the environment. From this model, potential tripping-hazard regions are automatically detected. Finally, the tripping risk of detected hazard region is quantitatively evaluated based on a combined human behavior and Monte Carlo simulation. The simulated tripping risk is compared with that measured using real humans, and it is revealed that the dimensional limit of risk evaluation accuracy in terms of barrier height on a walking terrain is very close to 10 mm.