Smart Testing for Smart Vehicles: Conditional Regression Strategies for Software Updates

Modern vehicles are software-defined, providing functionality, ensuring safety, and adhering to evolving regulatory standards. Consequently, software updates have become more frequent and essential, not only for fixing bugs and vulnerabilities but also for adapting to changing environments, introducing new features, and enhancing the user experience. However, each update carries the risk of unintended side effects, making thorough testing a critical component of the development and deployment process. System-level tests, in particular, are often resource-intensive and time-consuming. Saving testing time when safety issues arise is essential. To meet tight Time-to-Market (TTM) constraints and maintain development efficiency, it is vital to identify and execute only those tests relevant to the modified components of the system. Yet, in safety-critical domains such as automotive software, selective testing is permitted only under well-defined conditions to ensure continued compliance with safety standards. This paper investigates the conditions under which regression testing strategies can be applied safely and effectively in the context of automotive software updates. ... mehrIt categorizes different types of updates-corrective, preventive, adaptive, additive, and perfective-and analyzes their impact on development and testing artifacts. Based on this analysis, the paper proposes a classification framework that helps test engineers evaluate update scenarios and determine the appropriate subset of regression tests to execute, balancing efficiency with safety.