Applying 7 Failure Modes and STPA on Autonomous Systems: A Comparative Study

Safety analysis of modern autonomous systems is complex due to evolving software components and intricate hardware interactions. Traditional safety techniques like FMEA and FTA are not designed to cope with the new challenges raised by autonomous systems, e.g., their risk evaluation procedures cannot be used for the AI algorithms the systems are built on. In this paper, we report on promising safety analysis techniques, such as Seven Failure Modes (7FM) and Systems Theoretic Process Analysis (STPA). We compare 7FM and STPA, focusing on their ability to address complex system interactions and component-based failures. Our study examines a use case involving autonomous vehicle components to determine the effectiveness of each method in identifying and mitigating safety risks. Both techniques offer structured guidance for safety analysis but vary in approach and application. Our concluding remarks underscore the importance of integrating these methodologies for effective hazard analysis. We also discuss challenges encountered during the study, proposing improvements and future work to refine the safety analysis process.