The focus of this project is the modification of a commercially available in-pavement warning signal that was evolved from one originally designed to indicate the presence of pedestrians in a crosswalk. We have proposed use of a similar device to provide warning to vehicles approaching a railroad grade crossing, and we have tested a variety of illumination patterns in order to provide an optima...

The use of metadata expands on the opportunities for interoperability. By interoperability, it is meant the ability to develop conventions that enable data exchange and integration. Metadata descriptions from different domains are not semantically distinct but overlap and relate to each other in complex ways. As the number, size and complexity of the metadata standards grow, the task of facilit...

Crossing an urban traffic intersection is one of the most dangerous activities of a blind or visually impaired person's travel. Building on past work by the authors on the issue of proper alignment with the crosswalk, this paper addresses the complementary issue of knowing when it is time to cross. We describe a prototype portable system that alerts the user in real time once the Walk light is ...

Many cities are developing multiuse urban greenway trails to be used for recreation, exercise, and transportation. Analysts need many kinds of data about these new trails, but especially trail traffic because of its implications for the efficient allocation of resources for trail management. This paper addresses this need by presenting new information about the use of trails. We adapt and test ...

We propose an extension to the stateof-the-art Faster R-CNN detection model for multimodal pedestrian detection from RGB-D images. The proposed architectures address this problem by fusing convolutional neural network (CNN) representations. We elaborate two architectures, which primarily differ in the position of the fusion inside the model, and further compare several static and parametrized f...