Estimation of annual average daily traffic with optimal adjustment factors

This study aimed to estimate the annual average daily traffic in inter-urban networks determining the best correlation (affinity) between the short period traffic counts and permanent traffic counters. A bi-level optimisation problem is proposed in which an agent in an upper level prefixes the affinities between short period traffic counts and permanent traffic counters stations and looks to minimise the annual average daily traffic calculation error while, in a lower level, an origin–destination (O–D) trip matrix estimation problem from traffic counts is solved. The proposed model is tested over the well-known Sioux-Falls network and applied to a real case of Cantabria (Spain) regional road network. The importance of determining appropriate affinity and the effect of localisation of permanent traffic counters stations are discussed.