Detecting and counting pedestrians is a common task in transportation and it was among the first where the research team applied artificial intelligence technologies. Counting spaced individuals can be done very inexpensively with passive infrared (sensing the heat of a nearby object) but this breaks down when pedestrians are closely spaced and their path is not well defined. With AI, an entire image can be scanned for pedestrians and multiple images can employ a tracking concept.
The following is an example of working with a group of pedestrians on and near a sidewalk.
The work also included looking for bicycles and skateboards along the sidewalk and into the bike lane.