The purpose of this study was to rank the most relevant defects found on the roads managed by two companies from the road management sector located in the states of Minas Gerais and Bahia. To this end, questionnaires were conducted with engineers and professors from the forest road area and data were collected on the amount and intensity of sevendefects(inadequate cross sections, inappropriate lateral drainage, corrugation, excess dust, potholes, wheel track and aggregate loss), which were analyzed by the Unpaved Roads Condition Index (URCI) method. Aiming to provide greater accuracy on the assessment of the problems, Artificial Neural Network(ANN) techniques were employed in order to rankthe most serious defects thatcan make roads closed to traffic. Results show that 33% of the sections analyzed must undergo maintenance work, once they have reached the 'very poor' classification, which can cause hindrance to road users. Inadequate cross sectionsand inappropriate lateral drainagewere the most important defects found on the roads studied.
hierarchy; defects; roads maintenance