Automated Vehicles for All (AVA)
About the AVA Project
The Automated Vehicles for All (AVA) project brings together researchers across the United States to address challenges associated with deploying automated vehicles in complex transportation environments. Funded through the U.S. Department of Transportation's Automated Driving Systems Demonstration Grant Program, the multi-university project is led by the Texas A&M Engineering Experiment Station, in collaboration with The George Washington University, the University of California, Davis, and the University of Illinois Urbana-Champaign. At GW, the project is helping establish a research environment where autonomous-vehicle technologies can be developed and evaluated from simulation and behavioral studies to full-scale vehicle experimentation.
Goal and Objectives
At The George Washington University (GWU), the research focuses on developing and evaluating connected and autonomous vehicle technologies with an emphasis on vehicle safety, automated driving behavior, human interaction, decision-making, and real-world experimentation.
Key Research Objectives
- Autonomous Vehicle Perception and Decision-Making
Develop and evaluate methods that enable automated vehicles to understand complex driving environments and make appropriate driving decisions. - Vehicle Planning and Control
Develop vehicle planning and control approaches that translate autonomous-driving decisions into safe and reliable vehicle motion. - Human–Autonomous Vehicle Interaction
Investigate how automated vehicles interact with pedestrians, drivers, cyclists, and other road users in realistic transportation environments. - Ethical Decision-Making
Develop frameworks for incorporating human preferences, safety considerations, and ethical principles into autonomous-vehicle decision-making. - Real-World Testing and Evaluation
Use instrumented vehicles, roadside sensing, communication systems, and experimental environments to evaluate autonomous-driving technologies beyond simulation.
GW Research and Demonstration Platform
The George Washington University is developing a real-world research and demonstration environment for connected and autonomous vehicles at its Foggy Bottom campus. The research team has instrumented autonomous vehicles, deployed sensing infrastructure on university buildings and roadway segments, and developed communication and monitoring capabilities for studying vehicle operation in controlled and naturalistic environments
This environment allows researchers to investigate how automated vehicles:
- perceive surrounding transportation environments;
- interpret information from sensors and infrastructure;
- make driving decisions;
- interact with pedestrians and other vehicles;
- translate planning decisions into vehicle motion; and
- respond to the uncertainty and complexity of real-world transportation systems.
The platform is being used to connect simulation, behavioral modeling, vehicle automation, sensing, and physical vehicle experiments.
Research Areas
1. Autonomous Vehicle Planning and Control
GW researchers are developing and evaluating planning and control approaches for autonomous vehicles using a combination of mathematical modeling, simulation, sensing systems, and full-scale vehicle experiments. A major objective is to understand how driving models developed in theoretical or simulation environments behave when connected to real perception systems and physical vehicle actuators.
2. Ethical Decision-Making for Autonomous Vehicles
Automated vehicles may encounter situations in which safety, mobility, efficiency, and the interests of different road users must be considered simultaneously. GW researchers are developing data-driven frameworks for studying how ethical considerations can be incorporated into autonomous-vehicle decision-making.
3. Human Behavior and Autonomous Vehicles
Understanding human behavior is an important component of developing automated vehicles that will operate alongside human drivers, pedestrians, and other road users. GW researchers are investigating the relationship between what people say an automated vehicle should do and how people actually behave when they encounter similar situations themselves.