Testing automated driving systems (ADS) is complex as ADS exhibit multidimensional and large input spaces. A promising approach is search-based testing (SBT) based on simulation software, as it is more effective and less time-consuming than on-road testing. However, creating automated testing pipelines for ADS is particularly challenging, as it requires integrating complex driving simulation platforms and establishing communication protocols and Application Programming Interfaces (APIs) with the desired search algorithm. This complexity prevents the widespread adoption of SBT for ADS.

OpenSBT facilitates search-based testing by providing a modular and extensible open-source code base that simplifies the integration of search algorithms, fitness and criticality functions, and simulation environments. It offers the following functionalities:

• It allows the application of existing or user-defined search algorithms for testing ADS and the definition of fitness and criticality functions,
• It provides interfaces for integrating different Systems Under Test (SUTs) and simulators,
• It visualizes and analyses test results and supports the assessment of critical system behaviour.

Through its flexible architecture, OpenSBT supports the development of new testing methods as well as the systematic analysis of critical driving scenarios in simulation-based environments.

OpenSBT is open source and available for both academic and commercial use. Furthermore, OpenSBT has been evaluated by the automotive industry partner DENSO.