More than 1 billion motor vehicles travel roadways worldwide and that number is expected to double within one or two decades. The rise in automobiles and trucks have led to increased social, economic, and safety issues, such as the more than 30,000 deaths from car related accidents on US highways every year, according to a US Department of Transportation (DOT) report. Efficiency is also affected – a 2017 INRIX Global Traffic report found Los Angeles leads as the most congested urban area in the United States by spending over 100 hours of total drive time in congestion, a total cost of $19.2 billion for the city. To combat these increased social and economic issues of the motor vehicle rise, alternative methods to increase efficiency and sustainability of highway travel are being developed.
Ziran Wang, recent mechanical engineering Ph.D. graduate, and Xishun Liao, electrical and computer engineering Ph.D. student, built an immersive driving simulation platform that develops a cooperative ramp merging system. Highway on-ramp merging has been linked to traffic congestion. Drivers along the on-ramp adjust vehicle speeds and positions to enter the highway, while drivers on the highway accommodate vehicle speeds and positions to avoid collision with merging vehicles from the on-ramp, heavily affecting upstream traffic flows.
“Merging onto a highway is like dancing the tango while blindfolded – you’re negotiating who slows and who doesn’t while avoiding stepping on one another’s toes. But what if we could remove the blindfold and simplify the merging process?” said Wang.
In congested traffic conditions, such maneuvers if inefficiently performed will lead to high risks of accidents and excessive energy consumption and pollutant emissions. Under the direction of CE-CERT Director Matthew Barth and researcher Guoyuan Wu, Wang developed an innovative distributed consensus for Connected and Automated Vehicles (CAV) to cooperate with each other by Vehicle-to-everything (V2X) communications.
Agent-based modeling and simulation (ABMS) is a popular approach for modeling autonomous and interacting agents in a multi-agent system. Specifically, ABMS can be applied to connected and automated vehicles (CAVs) since CAVs can operate autonomously with the help of onboard sensors, and cooperate with each other through vehicle-to-everything (V2X) communications. Connected vehicles utilize different communication technologies to communicate with other vehicles on the road, roadside infrastructure, and the “Cloud”. The two main methods of communication are Dedicated Short Range Communications (DSRC), in which local Wi-Fi communication uses a specific channel to transmit their vehicle to vehicle (V2V) communication, and Accelerated Network, which collects data from a sim card attached to the vehicle and sends the data to the cloud.
Wang’s simulation adopts the Unity game engine to conduct ABMS of CAVs in a case study of cooperative on-ramp merging. His team developed an online feedforward/feedback longitudinal controller for CAVs to cooperatively merge at ramps. The vehicles can share information that allows them to adjust their speed before actually seeing the other vehicle. Merging behaviors are much smoother and safer, reducing fuel consumption and reducing travel time. This platform will be made open-source and be uploaded online for other research entities to use in the future.
“The simulation is beneficial for the entire transportation system, especially when you have these potentially dangerous ramp merging incidents, where the driver underestimates the distance and time needed to merge,” said Wang. “It can be difficult for human drivers to make decisions in just seconds. Imagine the benefits if this system were embedded in every vehicle.”
The simulation takes place on California Highway 237 in the City of Mountain View in Silicon Valley. Wang and Liao also tested the system on Columbia Avenue in the City of Riverside, which mirrors a similar demo in comparison to the game. Due to the need for vehicles to work together, system penetration requires every vehicle to be a CAV, but the team is working on broader alternatives.
In addition to testing the simulation with all vehicles being controlled by the proposed online feedforward/feedback longitudinal controller, Wang and his team conducted human-in-the-loop simulations where the merging vehicle is controlled by a human driver on a simulator. Compared to the human-in-the-loop scenario, the cooperative merging protocol provides an additional 7.8 percent savings on energy consumption and up to 58.4 percent reduction on pollutant emissions.
Results from this research have been published recently on Society of Automotive Engineers’ International Journal of Connected and Automated Vehicles. Due to his progress on this cooperative ramp merging system, Wang will begin work in July at the Toyota InfoTech Labs in Silicon Valley to perform similar research.