Rapid and Global Bridge Assessment for the Military

Abstract

Prior to any military operation, a critical task of military engineers involves the ability to quickly assess the load-carrying capacity of bridges. These assessments are required to facilitate the selection of movement corridors for troops, equipment, and supplies. Assessments are complicated when bridges are located in areas that are inaccesible or where information regarding design standards is unobtainable. The U.S. Army Engineer Research and Development Center has developed a systematic methodology to provide rapid, accurate bridge assessments on a large scale and in all regions. The methodology uses a machine learning approach designed to discover regional construction and condition tendencies given a sample of onsite inspections. Learning occurs based on the notion of bridge similarity. This approach allows for the completion of bridge assessments in a timely and effective manner. A brief description of the systematic methodology is provided, and the results and analysis of this approach are presented using an actual case study that illustrates the effectiveness of the approach. The results indicate that construction tendencies can be captured and applied to reduce the data collection effort while improving bridge assessments.