News | METRANS Researchers Ioannou & Monteiro Shape the Future of Autonomous Vehicles

Stop the Video



by Christian Spielmann

Autonomous vehicle technology is rapidly progressing, with many leading automotive manufacturers announcing their intent to bring vehicles that allow for autonomous driving to the market in upcoming years [1]. One major motivation in introducing autonomous vehicles to the general public is a projected reduction of fatal traffic accidents [2]. The U.S. Department of Transportation has found 94% of all fatal crashes to be caused by human error, meaning that there is potential to significantly reduce this share of fatalities. While increased vehicular safety is a motivator, ensuring safety of autonomous vehicles also represents a technological challenge. This is the primary reason why functionalities like autonomous lane changes are not yet offered by most established automotive manufacturers [3]. Changing lanes automatically is still an open challenge, especially in dense traffic. As part of the effort to overcome this challenge, METRANS researchers Petros Ioannou and Fernando V. Monteiro from University of Southern California (USC), conducted a research project examining how “vehicles communicate with each other and negotiate the creation of safe spacings in order to merge without taking any safety risks.”

Funded by the Pacific Southwest Region UTC, Ioannou and Monteiro’s project, “Connected Autonomous Vehicles: Safety During Merging and Lane Change and Impact on Traffic Flow,” focused on improving passenger safety and highway traffic flow in merging and lane change scenarios through communication between autonomous vehicles. In past iterations of autonomous vehicles, advanced driver assistance systems and autonomous driving systems had their actions dictated by monitoring their surroundings with sensors. However, this approach may not be sufficient to safely adapt to all traffic-merging scenarios efficiently. Additionally, some form of communication among vehicles is needed.


As part of the research project, a control policy has been developed that enables vehicles to manage complex lane change scenarios safely and smoothly with the help of electrical communication. Extensive simulation verifies and demonstrates the benefits of this policy that represents a new achievement on the road towards connected autonomous vehicles and the future of mobility.


To better understand this issue and the importance of connectivity for autonomous vehicles, let us consider a hypothetical situation. Imagine driving on a highway with dense traffic and you are forced to merge into the left lane to stay on the highway. A human driver would signal this upcoming lane change to other drivers, and another driver would respond accordingly by creating a gap in traffic that allows you to merge in. However, if all vehicles on your left are autonomous driving vehicles, how will they be able to sense your request and act accordingly? There must be some form of intercommunication. Part of the research project includes a control policy that was developed to enable vehicles to manage complex lane change scenarios safely and smoothly with the help of electronic communication. This allows the autonomous vehicles to generate the desired “smooth gap generation behavior” which will allow the cars to safely and seamlessly merge, even in heavy traffic scenarios. Extensive simulation tests verify and demonstrate the benefits of this policy. It represents a new achievement on the road towards connected autonomous vehicles and the future of mobility.




[1] Daimler. “Mercedes-Benz and NVIDIA: Software-Defined Computing Architecture for Automated Driving Across Future Fleet.” Daimler. Accessed June 29, 2020.


[2] National Highway Traffic Safety Administration. “Automated Driving Systems: A Vision for Safety,” U.S. Department of Transportation. Accessed June 29, 2020.


[3] Rachel Siegel. “Tesla’s Automatic Lane-Changing Feature Is ‘Far Less Competent’ than a Human Driver, Consumer Reports Says.” Washington Post. Accessed June 29, 2020.