The Development of Safety Cases for an Autonomous Vehicle: A Comparative Study on Different Methods

摘要

The Connected and Autonomous Vehicles (CAVs) promise huge economic, social and environmental benefits. The autonomous vehicles supposed to be safer than human drivers. However, the advanced systems and complex levels of automation could also bring accidents by tiny faults of hardware or errors of software. To achieve complete safety, a safety case providing guidance on the identification and classification of hazardous events, and the minimization of these risks needs to be developed throughout the entire development lifecycle process of CAVs. A comprehensible and valid safety case has to employ appropriate safety approaches complying with the automotive functional safety requirements in ISO 26262. The technical focus of present work is on the comparative study of different safety approaches, in particular, Failure Mode and Effects Analysis (FMEA) method and Goal Structuring Notation (GSN) method that have been employed to generate lists of hazardous events, safety goals and functional safety requirements at the vehicle level. A case study on the safety case development of INISIGHT autonomous vehicle has been carried out using the aforementioned methods. This case study covers the safety argument of battery and charging system that supply the whole electric power for INSIGHT vehicle. The safety of this systems has been assessed along with their potential for malfunction together with the layers of protection. The results and conclusions from case study analyses suggest the safety case of CAVs can be developed in a highly effective manner by employing a combined method of GSN and FMEA.