Comparing the effect of SAE-2 and SAE-4 equipped vehicles on stop-and-go waves at bottlenecks

Abstract

Stop-and-go waves on highways are a product of human acting. As humans we are not perfect and other drivers are often hampered when changing lanes or braking. These disruptive actions can be prevented by introducing autonomous vehicles in the vehicle mix. In this research self-driving vehicles are represented by SAE-4 level vehicles. Autonomy levels of vehicles are made by the Society of Automotive Engineers (2014) and range from 0 to 5. SAE-0 are human-driven, while SAE-5 would correspond with a vehicle able to handle every situation. In this research human-driven, high-level automation (SAE-4) and low-level automation (SAE-2) are studied.

By creating a model which can simulate human drivers, low- and high-level automation on a highway, different vehicle mixes are explored. These vehicle groups are different in time headway and acceleration and are spawned on the main lane and an incoming on-ramp. In the base scenario with 100% human-driven vehicles the on-ramp location is where the disruptive manoeuvres happen. Due to the high amount of vehicles on both lanes stop-and-wave forming actions are inevitable. These waves propagate with 7 km/h upstream and cause an average speed of 96.9 km/h which is 23 km/h lower than the maximum speed of 120 kmh. Besides the lower average speed is 1% of time spent in uncomfortable accelerations which decrease safety. The capacity of the base scenario is 3600 vehicles per hour. If SAE-2 vehicles are added to the mix the capacity drops as does the average speed. In a scenario where 25% and 3% are equipped with SAE-2 and SAE-4 respectively which may happen in 2030, a capacity drop of 11.9% is observed. In addition the average speed also decreases with 10 km/h. At the same time the amount of uncomfortable deaccelerations decreases as well. In this scenario comfort and safety are increased while traffic flow is significantly worsened. This pattern is also visible in the 2035 and 100% SAE-2 scenarios, albeit that the effects are felt harder in these cases.

In contrast with the SAE-2 vehicles are SAE-4 equipped vehicles able to improve the traffic flow conditions. When 18% or more of the vehicles is equipped with self-driving capacities positive effects for traffic flow are felt. Capacity increases and less slowdowns are observed. If all vehicles would be autonomous, capacity would rise to 6000 vehicles per hour, congestion isn’t present and all vehicles can drive the maximum set speed.