Reengineering of Simulation-based Digital Twin Toolchains for Automated Vehicles to Mitigate Technical Debt: A Case Study

In early vehicle development, simulation-based Digital Twin toolchains enable simulated testing of automated driving functions prior to real-world testing. In a research setting, they are created by researchers and comprise research software prototypes, proprietary tools, and customized open-source software. Although this composition initially yields quick results, it accumulates technical debt and erodes reproducibility. We study a case of architectural technical debt in simulation-based Digital Twin toolchains for automated vehicles and a) identify symptoms of architectural technical debt present in the toolchains via analyses and interviews with researchers, b) investigate the underlying architectural technical debt causing these symptoms, c) relate them with categories in the field of architectural technical debt, d) reengineer the toolchains towards a containerized service-based architecture, and e) discuss the tradeoffs and lessons learned. Our results show that an intermediate implementation already substantially mitigates architectural technical debt while preserving performance metrics.