摘要：AbstractAnnual Average Daily Traffic (AADT) is one of the most important pieces of data being used widely in planning, design, operation and management of roads and facilities. A reliable estimate of AADT has always been one of the main interests of transportation and highway agencies. Traditionally, AADTs are estimated for the majority of road segments of a network from short-term traffic counts (STTCs or coverage counts) by applying a set of expansion factors derived from permanent traffic counts (PTCs). Literature indicates that the FHWA (Federal Highway Administration) method may be most widely used. In this method, roads in the same functional class are assumed to have similar traffic patterns, and the factors derived from PTC data collected from the class is used to convert STTCs to AADT estimates. However, it should be noted that, because roads from a functional class do not necessarily have similar seasonal traffic variations, this method may sometimes produce large estimation errors. In this regard, this paper proposes a novel pattern-matching method, which constructs a seasonal traffic variation profile for a short-term counting site using all historical counts available and then use this profile to assign the site to a PTC or a PTC group. In addition, a Bayesian approach is developed to explicitly consider and show the “risk or uncertainty” associated with assigning each short-term counting site to different PTC or PTC groups. Study results based on the simulated STTCs from a permanent counter on a winter recreational road in Alberta, Canada show that the new method proposed can limit the 95th percentile (P95) AADT estimation error to less than 13%, in contrast to 21.7% from the FHWA method. Moreover, it should be noted that the proposed method will not impose any additional monitoring cost, or make any change to existing traffic monitoring programs, and therefore it will be easy for highway agencies to incorporate the proposed method in their practice.
