Development of an Algorithm for Solving the Problem of Determining the Size of the Deflection of Road Barriers Based on Video Recordings of Full-Scale Crash Tests

Abstract

The most important state task is to ensure road safety. The solution is achieved due to various factors, the most important of which is the development of road transport infrastructure, the use of modern designs of road facilities and, first of all, road barriers, the purpose of which is to prevent road accidents and reduce the number and severity of injuries of injured participants in an accident. The design of modern road barrier structures requires the use of innovative digital design and analysis methods, as well as artificial intelligence methods for decision making at all stages of the life cycle of these products. The article discusses the use of computer vision and machine learning methods for the refined determination of the values of dynamic and residual deflection of the road barrier based on the results of video and photography of the field test process, which are carried out by driving a vehicle onto the tested road barrier. The obtained refined parameters will improve the quality of validation for subsequent virtual tests that are carried out for similar brands of modified fence designs, more accurately assess the effectiveness of various fence designs. Thus, the choice of the optimal design of the road barrier will be decided with lower resource costs and with a high degree of compliance with the specified requirements of the road transport infrastructure.