Systematic Literature Review of Corner Case Identification and Generation

This ongoing systematic literature review examines the current state of corner case identification and generation for autonomous vehicle (AV) testing, emphasizing the challenges and opportunities in advancing safety and robustness. Corner cases, representing rare but critical scenarios, are crucial for assessing AV systems’ performance under extreme or unexpected conditions. Existing reviews have explored related areas such as software verification, risk assessment, and driving system testing but often fall short in targeting corner case generation or addressing practical applications due to outdated scopes, limited focus, or reproducibility issues.

This work aims to build upon and extend previous studies by synthesizing recent developments in corner case methodologies, incorporating broader regional analyses, and exploring the application of these cases to enhance AV safety. The review categorizes the literature into identification-focused, generation-focused, and hybrid approaches, offering insights into the interplay between these methods and their roles in testing and validating AV systems.

As part of this effort, the review seeks to identify key challenges, including the lack of standardized frameworks, limited geographical diversity in research, and gaps in reproducible methodologies. By addressing these issues, this ongoing study aims to provide a roadmap for future research and to foster interdisciplinary collaboration, ultimately advancing the field toward safer and more reliable autonomous vehicles capable of handling real-world complexities with greater confidence.