This paper analyses neural network use cases for road tolls. Multiple road toll collection methods are analyzed in this paper: vingettes, toll platforms, toll gates and all of these analyzed toll collection methods have one of the main problems - if only some of the vehicle categories are taxed for road usage, vehicles have to be classified. This paper proposes to use neural networks to determine road tolls, because they solve main problems that standart road tolls collection methods face. This paper describes the analysis of vehicle classification using neural networks. Several different methods are analyzed: photos, magnetic sensors and inductive loops. Vehicle photo classification and a neural network model, based on experiments described in this paper, is proposed. This classification method has been chosen because, compared to other analyzed methods, it classifies vehicles into more classes, the accuracy of the model does not fall behind other methods, and the installation of the equipment is simpler.
Transporto priemonių kelių mokesčių nustatymas naudojant neuroninius tinklus ; Determination of road taxes for vehicles by neural networks
2020-06-10
Hochschulschrift
Elektronische Ressource
Lithuanian , Englisch
DDC: | 629 |