Write a Blog >>
ICSE 2023
Sun 14 - Sat 20 May 2023 Melbourne, Australia

DoppelTest is a Python framework implemented to evaluate a novel autonomous driving software (ADS) testing approach discussed in the paper titled Doppelganger Test Generation for Revealing Bugs in Autonomous Driving Software. This research artifact implements a framework that orchestrates multiple instances of the same ADS and generates virtual scenarios with those instances. Since all vehicles in the virtual scenario are controlled by different instances of the ADS under test, any actual violation that occurs by or among them inherently reveals ADS misbehavior, thus revealing ADS bugs. The reviewer is assumed to have basic knowledge of the Ubuntu operating system, running bash commands, running Python scripts, or using an Oracle VM VirtualBox image. We provide detailed documentation on how to install the ADS and DoppelTest using a machine running Ubuntu 18.04 or above, or alternatively, the reviewer may use the provided Oracle VM VirtualBox image. Because of the nature of the framework, i.e., running multiple instances of ADS simultaneously, we recommend using a machine with at least 64 GB memory when evaluating DoppelTest. The badges this artifact is applying for are Reusable and Available.