Enabling Efficient Automated Vehicle Operations via Vehicle-to-Infrastructure Communications

Automated vehicle (AV) development and testing is well underway, but the majority of this testing is being conducted in dry and benign weather conditions. The vehicle perception systems, including cameras, lidar, and radar, are critical for edge detection, lane-striping observation, and detection of obstacles, but are negatively impacted by adverse weather conditions. Precipitation can decrease object detection by over 15%, and fog may decrease object detection by as much as 50%. Additionally, many of these state-of-the-art AVs are manufactured as electric vehicles (AEVs), and weather directly impacts AEV range in addition to the AV instrumentation and automated driving systems providing a burden on the vehicle’s battery.

Addressing this energy concern is at the core of the Department of Energy sponsored Infrastructure Enabled Energy Efficient Autonomous Vehicles project, which seeks to address the computational load on the AEV through offloading computational processing to the infrastructure. With weather providing a hazard not only to the maneuverability of a vehicle on the roadway, but also the perception system of an AEV, the ability to provide good weather observations and information to the vehicle itself is an important part of the project. An AEV was driven through multiple types of weather conditions and weather data collected via a vehicle-mounted instrument. Impactful weather conditions, including precipitation, road condition, air temperature, etc., were analyzed between the on-board vehicle instrument and the infrastructure to determine whether or not the existing infrastructure is sufficient for operations, or if observation at the vehicle is necessary.