Modern embedded and autonomous systems are pushing the boundaries of software complexity, especially in critical applications. Traditional testing methods often struggle to meet the demands of these systems, particularly when operating on resource-sharing architectures running complex operating systems like Linux. To address this challenge, we introduce Statistical Path Coverage (SPC), a novel statistical approach designed to enhance test effectiveness by statistically focusing on the execution paths exercised by target applications.

This presentation will discuss how SPC can quantify execution path coverage, estimate the risk of untested paths, and support assurance. We will also introduce DB4SIL, a tool leveraging FTrace to collect and analyze execution traces, enabling actionable insights into the kernel’s behavior during testing campaigns. Through examples, we will demonstrate how SPC and DB4SIL can guide developers in prioritizing testing efforts, improving test coverage, enabling continuous monitoring, and reducing risk in complex, software-driven systems.