Combining Semi-formal and Formal Methods for Safety Control in Autonomous Mobility-on-Demand Systems

Abstract

Ensuring the safety control of Autonomous Mobility-on-Demand systems is one of the biggest challenges facing designers to successful deployment. The addition of adaptability to such systems further hardens and delays modelling and validating phase, especially due to the current lack of design models and tools. The formal methods have proven to be useful for making the development process reliable at early design stages. Based on this approach, this paper proposes a mixed process to specify, design and verify safety requirements in adaptive AMoD Systems. This process provides analytical proofs of safety requirements during the design stage of a system when changes are cheap. This contribution deals with combining the UML MARTE profile for modelling the workload behaviour of the system and the formalism Net Condition Event System for consistency validation of safety properties. To verify the effectiveness of our proposal, several formal analyses are carried out using the model checker SESA. The evaluation of the proposed architecture, simulated by the Sumo software, proves the impact of the number of autonomous vehicles on the global performance and the intended quality of service (QoS) in the framework of the TORNADO project.