The majority of lethal train accidents occur at grade crossings, where the rail intersects with a public road or street. Many of these accidents are the result of deliberate action on the part of pedestrians or drivers who ignore signals and even evade barriers placed in their way. Eliminating the problem by eliminating the grade crossings is far too expensive, as creating an over or underpass costs millions of dollars and there are hundreds of thousands of grade crossings.

The alternative is to find a way to warn the train and/or authorities when something is blocking the track or remains in the area of the track too long. Initially with the support of the New York State Energy Research and Development Authority, and now under contract to the Federal Railway Administration, IEM developed a smart-video based method for detecting, tracking, and identifying potential intrusions into the grade-crossing region.

The initial work led to, and later work leverages from, our NASA project for aircraft security that resulted in our U.S. Patent 7,355,508 (System and Method for Monitoring an Area). The Smart Grade Crossing Monitor uses multisensor fusion to perform reliable detection of targets and discriminate between actual targets of interest (people, bicycles, cars, etc.) and distracters (small animals, blowing papers, shadows, and so on). Testing of the current prototype on a real roadway with simulated railway intersection demonstrated that the combined approach missed absolutely no intrusions by relevant targets, even though the analysis of the individual methods could, and did, miss some intrusions.

Development of the Smart Grade Crossing Monitor continues, and IEM expects to demonstrate and begin the marketing of this new and exciting product in the next few months.