MECHANISM OF ROAD CAPACITY UNDER DIFFERENT PENETRATION SCENARIOS OF AUTONOMOUS VEHICLES

Abstract

This paper takes L0 (ordinary vehicles), L2 and L4 (autonomous vehicles) as the research objects, adopts different car-following rules and improves them respectively based on driver's personality factors and variable time headway (VTH) strategy, introduces a benefit parameters to distinguish lane changing ability of them, and evaluates road capacity under mixed traffic flow with a basic diagram model and average travel time. Use SUMO to build a simulation platform and conduct a real-time systematic research based on Python. The results prove that: (1) After the penetration rate of autonomous exceeds 60 %, road capacity can be effectively improved, and the maximum increase of 32.52 % occurs in 100 % penetration scenario. (2) When traffic density is less than 27 vehicles/km, the average speed continues to be the maximum in 100 % scenario, and when it is greater than 27 vehicles/km, the critical penetration scenario is 80 %. (3) The average travel time begins to decrease after the penetration rate exceeds 20 %, and can be reduced by 23.38 % in 100 % scenario. It shows that traffic efficiency is closely associated with penetration rates of autonomous vehicles.