Non-Deterministic Vehicle Simulation Tool

Safety of autonomous vehicles is a pressing concern in the automobile industry. One core component of autonomous driving safety is for each vehicle to perform a safety analysis based on the current state of other vehicles on the same road. Given well-defined vehicle models, state transitions and a road definition, an autonomous agent can analyze its own safety in a traffic scenario for the next T units of time. However, existing autonomous driving simulation tools rely on deterministic behavior of agents on fixed set of roads. In this thesis, we develop a software tool with Go for performing non-deterministic simulations of traffic scenarios on user-defined roads. Additionally, our tool will render visualizations that traverse the entire simulation tree, which we will then use to perform vehicle safety analysis in real time. In the end, we will conclude with a plotting tool that visualizes reachtubes of all the vehicles in the simulation.U of I OnlyUndergraduate senior thesis not recommended for open acces