Autonomous vehicle impact on improving road network vulnerability

Abstract

Abstract Purpose This study first presents a method to identify the parameters increasing road vulnerability on a macroscopic road network model. The second part exploresthe effect size difference of the analyzed attributes on network vulnerability through the implementation of different autonomous vehicles (AVs) penetrations and automation levels. Methods The road traffic network of Budapest, Hungary on PTV VISUM is studied by adopting a passenger car unit factor procedure to simulate the effect of AVs on road saturation. Five link parameters were used: length, distance from the centre, speed, number of lanes, and number of connectors. Network vulnerability was studied by simulating a combination of road elimination process with different passenger car unit values for AVs. Results The analysis found the number of road lanes is the most significant parameter, affecting the link criticality; followed by road length and distance from the centre. The analysis of four AV scenarios with different AV penetration and level of automation showed huge effect differences ranging from 3.50% for a simple AV automation level with low AV percentage to as large to 28.53% for a fully automated fleet. Conclusions AV implementation has proved efficient in reducing the amount of travel delays in the case of road failure. Finally, it was found that the number of lanes remained the most significant influencing parameter on travel delay. The main question is to discover the effect size difference of the analyzed attributes on network vulnerability through the implementation of different AVs penetrations and automation levels.