In the last few years, an increasing amount of research has been dealing with the development of autonomous vehicles (AVs). With the increasing deployment of the AVs on the roads over the years it is necessary to adapt road infrastructure to accommodate them. This paper gives an insight into the impact of AVs on the road infrastructure such as horizontal and vertical alignment, and dimensions of cross section elements. The implementation of the AVs into the road network will not depend just on the preparation of road infrastructure but also on the future development of vehicle automation and the proportion of AVs according to conventional vehicles (CVs) into the vehicle fleet. Therefore, according to AVs proportion into the fleet, three different scenarios for road infrastructure upgrades are given. Finally, the advantages and disadvantages of the introduction of AVs regarding the requirements for existing and future road infrastructure are discussed.

Roundabout design is an iterative process consisting of a preliminary geometry design, geometry performance checks, and the estimation of intersection functionality (based on the results of analytical or regression models). Since both roundabout geometry design procedures and traffic characteristics vary around the world, the discussion on which functionality estimation model is more appropriate is ongoing. This research aims to reduce the uncertainty in decision-making during this final roundabout design stage. Its two objectives were to analyze and compare the results of roundabout performance estimations derived from one analytical and one regression model, and to quantify the model results’ susceptibility to changes in roundabout geometric parameters. For this, 60 four-legged single-lane roundabout schemes were created, varying in size and leg alignment. Their geometric parameters resulted from the assumption of their location in a suburban environment and chosen design vehicle swept path analysis. To compare the models’ results, the degree of saturation of roundabout entries was calculated based on presumed traffic flows. The results showed that the regression model estimates higher functionality and that this difference (both between the two models and regression models applied on different schemes) is more pronounced as the outer radius and angle between the legs increase.

The introduction of new technologies, such as artificial intelligence in cars, raises the question of what the road infrastructure should be like in the future. The changes are expected to be significant, given that the human factor has influenced a number of design parameters. The ability of autonomous vehicles to drive themselves, anticipate situations, communicate with surrounding vehicles and infrastructure, and the environment in which they are located, places new demands on road infrastructure. Gradual changes in infrastructure will mostly depend on the speed of development of autonomous vehicles and their introduction into the transport system. Starting with intensive maintenance of roads and accompanying facilities, through separate corridors only for autonomous vehicles, while the goal is simplified and safe road infrastructure. The aim of this paper is to provide a review of the literature for the possible adaptation of road infrastructure intended for autonomous vehicles and to lay the foundation for further research in this area.