Abstract: Advanced Driver Assistance Systems (ADAS) are deployed on a massive scale in the automotive industry today because they enable safer transportation for vehicle passengers. A major limitation of the current state of ADAS and Advanced Driving Systems (ADS) is that they are only capable of operation during a limited set of conditions, otherwise known as the Operation Design Domain (ODD). Once these systems reach a technical readiness level to achieve operation beyond current limited ODDs, they hold promise to enable more efficient driving behaviors by optimizing vehicle speeds through intersections, in traffic, and in other various driving scenarios. To enable a pathway for deploying optimized ADAS control maneuvers, a greater understanding of the environmental impacts on ADAS perception systems is necessary. Environmental impacts are caused by external disturbances to the ADAS sensors, such as intense glares, nighttime driving, or soiled sensors, however, the greatest contributor to external disturbances is atmospheric effects caused by adverse weather.

This project focuses on characterizing the impact of weather on sensors used in ADAS (cameras, radars, and LiDARs). To enable this characterization, researchers from the Transportation and Power Systems Division (TAPS) have begun working with Environmental Science Division (EVS) at Argonne to gather Present Weather Sensor data including metrics such as visibility, precipitation rate, droplet size, and precipitation type for a more objective evaluation of weather. At the conclusion of this year, we will have a quantitative comparison of LiDAR, camera, and radar object detection capabilities vs objective weather metrics allowing us to understand more about the disturbance transfer behavior of sensor data against quantified weather metrics.