Road departure is a leading cause of fatal crashes in the US and half of all the crashes are related to road departure [1]. Road departure warning (RDW) and road keeping assistance (RKA) are the new active safety areas to be explored. Most of the currently available road-departure detection technologies rely on the detection of lane markings, which are either missing or unclear in many roads. Therefore, in additional to the these lane markings, next-generation road departure detection should rely on the detection of other road edge and boundary objects. Common road edge and boundary indicators include lane marking, grass, curb, metal guardrail, concrete divider, traffic barrels and cones. This paper investigates the distribution of major types of road edges and road boundaries in the United States in order to enhance and evaluate the capabilities and effectiveness of RDW and RKA. The paper describes the road location sample used for the analysis, presents the percentage of various types roadside objects in terms of number of appearance locations, percentage miles (%miles), and percentage car-miles (%car-miles = %miles*car_density). The percentage of roads that do not have any lane marking and do not have clear lane marking is also described. The representative shapes of each type of roadside objects are studied. The result is applicable for the development and evaluation of road departure warning and road keeping assistance systems.
展开▼