The Federal Railway Administration track safety standards require rigorous visual inspections of tracks based on their operating speed (FRA track class). Often these inspections are carried out using hy-rail (highway/rail) vehicles, with the trained inspector using a set of hand tools (track level, string line, gauges, etc.) to further measure locations that appear to be out of compliance. Currently, bolt-on inspection systems for use on inspector’s hy-rail vehicles such as track geometry measurement systems, can be used to assist and supplement the inspector, but are quite expensive. This research developed a prototype, low-cost, “smart” hy-rail wheel (SmartWheel) for deploying on an inspector’s hy-rail vehicle (or any hy-rail vehicle the railway operates) that assists the inspector in identifying locations in track with certain classes of potential defects, in an autonomous and passive manner. It was intended that the SmartWheel be self-contained, autonomous, and provide alerts to the operator. Additionally, the SmartWheel needed to be inexpensive to implement and provide additional information to the inspector to assist in assessing particular elements of the track condition. The innovative approach utilized a low-cost inertial measurement unit (IMU) integrated into the hy-rail gear along with a combined mechanistic and artificial intelligence (AI) approach to analyze the response data from the IMU to identify particular classes of track defects (or issues). These included profile/surface, cross level, curvature, dipped joints, rail surface defects, rail corrugation, mud spots, etc. This approach does not require a sophisticated algorithm for transforming the IMU data to measurable geometry parameters (which requires additional expensive hardware).  Rather, the system evaluates the IMU response data directly using AI algorithms developed as part of this research. The current status of the product addresses a subset of track geometry parameters. The primary benefit of this product is a safer operating environment through the low-cost implementation of a tool that assists inspectors in an autonomous fashion in locating potential track defects. A secondary benefit is identifying locations with habitual problems where revised maintenance practices can increase safety and reduce overall costs. While not all track anomalies are identifiable through this technology, a significant number of safety related anomalies are identifiable.

The Final Report is available here.