Cooperative driving automation based safety applications for Trucks: data-driven framework

Abstract: With the innovations in connected transportation technologies (vehicle-to-everything), we are moving towards mixed traffic environment consisting of legacy human drivers, connected vehicles, and units with varying level of automation. The way we envision traffic safety is at a turning point where we need to account for the dynamic interaction among these three different classes of vehicles. Within the same context, Cooperative driving automation (CDA) has gained significant traction in the last decade. However, most research has been focused on the light duty vehicles and advanced safety applications are built with validation. To complement the efforts, we focus on the-Commercial motor vehicles, including trucks and buses, that are critical for our economy and supply chain stability. We developed a data-driven testing and evaluation framework by integrating software and hardware based tools--a data-driven hybrid framework. The hardware components include roadside units (RSU), on-board units, traffic controller box (e.g., Econolite ASC-3), and scaled truck. The software components has two different systems: first, the truck can be simulated within the CARLA environment (vehicle level simulation), and second, network simulation (e.g., SUMO or PTV Vissim) to assess the network effect. The framework will leverage the CARMA platform to utilize CDA capabilities. The framework will fuse data from synthetic environment (CARLA) and real-world data (trajectories and public sources). The safety applications will focus on preventing targeted crash types identified from public database such as FARS and state-level data sources (e.g., public CMV crash datasets available through public websites).