Empirical Studies on Traffic Flow in Inclement Weather
Figure ES.1. Weather Events and Traffic States Occur: First, weather events occur on a road network, affecting a particular traffic state. This image presents a storm cloud raining on a roadway. Measurement: Second, measurement devices measure aspects of the weather event and the traffic states. This image presents and ASOS/AWOS weather sensor and a loop detector. Goal: Accurately reproduce reality with imperfect measurement instruments. Issues: Data may require cleaning and processing. Data Fusion: Third, data from several weather measurement devices are matched with data from a single set of traffic measurement devices. This image presents a loop detector and the nearest ASOS weather sensor. Goal: Isolate weather events geographically and match to detectors. Weather Categorization: Fourth, weather data are placed into categories, using a combination of statistical techniques and intuition. This image presents raindrops and three levels of categories: Heavy rain, Light rain, and Fair. Goal: Categorize weather into statistically significant bins. Issues: Categorization must be useful to engineers. Traffic Analysis: Traffic data from differing weather categories are used to fit a model and estimate critical traffic stream parameters. The differences in model fit determine the impact of a weather category on traffic parameters. If needed, weather events are re-categorized until significant traffic parameter impacts are found. This image presents a plot of data points in a speed- flow curve. Goal: Determine impact of different weather categories on traffic parameters. Issues: May need to iterate back to categorization until meaningful traffic differences are found.
