A Review Study on Functional Safety Verification and Validation for Autonomous Vehicles

In recent years, autonomous vehicle studies have been rapidly increasing with technological developments in the automotive industry. While improving autonomous driving performances and increasing the diversity of autonomous vehicles, these studies bring new challenges as well. In this study, a review study on functional safety verification and validation for autonomous vehicles, which are developed using machine learning, neural networks, and artificial intelligence structures, is presented considering as one of the main challenges in this domain. Within the scope of this study, "What kind of difficulties exist in safety verification and validation processes", "Which theoretical and experimental methods are proposed against the stated difficulties", "Which software and hardware tools are used", "Which approaches are recommended for ISO 26262 compliance" and "how autonomy level changes the existing verification and validation methods" are investigated. A comprehensive literature review including highlights from papers is presented with respect to the key research questions.