Vehicular Traffic Modeling and Control in Mixed Manual and Automated Environments

University Of Wisconsin-Madison, Madison WI

Investigators

Abstract

Cutting-edge vehicle technologies such as Connected Vehicles and Automated Vehicles present unprecedented opportunities to drastically improve traffic operations and safety. These technologies can fundamentally change driver interactions and have enormous potential to remedy traffic phenomena known to be detrimental to traffic efficiency and stability. Among different types of Automated Vehicles technologies, Cooperative Adaptive Cruise Control is particularly advantageous due to its unique ability to foster high performance, doubling (or more) roadway capacity and significantly improving flow stability. This research project seeks to shed light on the traffic congestion mechanisms in mixed streams of manual and high performance automated vehicles. It also aims to develop mitigation strategies to reduce traffic congestion by unprecedented levels, contributing to the nation's economic competitiveness and sustainable urban development. This project will engage in a range of integrated research, educational and outreach activities that will expand the knowledge obtained from this research to a broader audience. The activities include (i) developing simulation-based educational modules, (ii) disseminating results using driving simulator, and (iii) engaging undergraduate and graduate students in the research and education. Traffic breakdown (onset of wide-spread congestion) is often triggered at freeway bottlenecks near merges and weaves. This phenomenon is characterized by high flow prior to breakdown, succeeded by a significant reduction in bottleneck discharge rate (known as "capacity drop"). The objectives of this research are to: (1) shed light on the behavioral mechanisms underlying traffic breakdown at bottlenecks in mixed manual and Cooperative Adaptive Cruise Control enabled vehicular environments and (2) develop theoretically-grounded control strategies to mitigate traffic breakdown and capacity drop. To better understand these mechanisms and enable vehicle-based control, this research will perform multi-scale analysis and modeling by linking microscopic features (driver characteristics, lane changes) to mesoscopic features (vehicle platooning) and eventually to macroscopic features (breakdown flow). This research will push the frontier of traffic flow research in the era of automated vehicles. Results from this research will advance our knowledge of congestion mechanisms, particularly around extended merge and weave bottlenecks. Moreover, control strategies will be developed based on the fundamental understanding of driving behavior to effectively mitigate traffic breakdown via vehicle-based control. Both proactive and reactive strategies will be developed at different levels of sophistication to enable implementation of robust control for select automation levels.