Field Evaluation of Automatic Pedestrian Detectors in Cold Temperatures

The equitable accommodation of pedestrians is critical in transportation engineering and planning. Major concerns over the use of push buttons to activate pedestrian signals have been raised in many jurisdictions because elderly pedestrians and certain physically impaired pedestrians experience difficulty using push buttons, even when the push button is placed in a convenient and standard location. This paper presents the results of an analysis of the performance of three commercially available curbside automatic pedestrian detectors (APD)—a passive infrared and stereovision curbside detector, a passive infrared curbside detector, and a microwave detector—in the field as a function of weather, temperature, and temporal variations at signalized intersections during the winter months at temperatures ranging from −34°C (−29°F) to 0°C (32°F). The results were classified according to detector sensitivity, which referred to the percentage of pedestrian crossings detected successfully, and detector selectivity, which referred to the percentage of activations triggered by actual pedestrians, instead of false activations from vehicle movement, trees in the wind, or other causes. From a sample of 8,225 detections at two sites, the results revealed that the three APDs were generally sensitive at sending valid pedestrian calls, but selectivity rates remained less than 50%. Under all conditions and of all APDs analyzed, the infrared detector consistently had the highest sensitivity and lowest selectivity; the infrared–video detector had the second-highest sensitivity and the highest selectivity; and the microwave detector had the lowest sensitivity and the second-highest selectivity.