A Domain Specific Modeling Language for Safe Autonomous Vehicle Behaviors with Dynamic Constraint Verification

Christopher Alicea-Nieves (Interamerican University of Puerto Rico, Bayamon Campus)
Kaitlyn Oura (University of Arizona)
William Silloway (Kennesaw State University)

Safety is one of the most important aspects of cyber-physical systems due that many interfaces have already been made available for experts and non-experts of the fields involved in these systems, allowing them to develop behaviors for these automated systems such as automated cars. These behaviors can constitute primitive movements, as well as complex ones. Experts in the field have been developing tools for the verification of dynamic behavioral constraints of a system that allow us to determine whether the behavior will be correct or not, and the generation of lower-level artifacts for said systems. With these tools, individuals and groups of both experts and non-experts in the field can program behaviors for these systems and ensure that their behaviors are correct and safe.In order to achieve this, a high-level domain-specific modeling language was created so that non-experts could design behaviors and run them on an automated vehicle. The operations that the vehicle can perform are limited, but it still allows for the vehicle to perform unsafely. The modeling language is compiled into code that can be run on a virtual machine that interprets the model's code and sends safe commands to thevehicle. Verification for the system was performed at a design-level which includes the constraints imposed in our modeling language and verification of the state machine of said behavior.