A Mobile Interface designed for Safety Verification of Autonomous Vehicle Trajectories

Researchers: Joanna De Los Santos, Tarif Haque, Duc Lam

Abstract: Safety verification of autonomous vehicle (AV) trajectories is a common and complex research problem. In this paper, we propose a mobile Android interface that facilitates safe interaction between a human and an autonomous vehicle. A user interface (UI) which allows for the specification of a trajectory for an AV to follow is presented. To verify the safety of human-input trajectories, a series of checks are performed on the specified path at runtime. The proposed system allows the user to be in control of the car, while delegating low-level control to the vehicle. MATLAB simulation techniques and tests performed on the Cognitive and Autonomous Test Vehicle (CATVehicle) at the University of Arizona are used to verify the generated trajectories.

Keywords: trajectory generation, runtime safety checks, autonomous vehicle safety verification, human-computer interaction, user interfaces