chapter  260
A simulation-based safety analysis framework for autonomous vehicles—assessing impacts on road transport system’s safety and efficiency
ByL.F. Vismari, C.B.S.T. Molina, J.B. Camargo, J.R. Almeida, R. Inam, E. Fersman, M.V. Marquezini
Pages 9

You have download access for this chapter.

DownloadPDF 1.09MB

Advances in Information and Communication Technologies (ICT), pervasive computing and Artificial Intelligence (AI) are affecting all daily human life domains in multiple ways. Safety-critical transport systems are being massively affected by this technological shift. The Autonomous Vehicles (AV) are being considered the most promising new element in this future transport paradigm. It is expected that the AVs bring relevant benefits to the society, mainly reducing accidents and increasing accessibility. Given that any new safety-critical system, concept or technology is placed in operations only if its benefits outweigh the safety risks, assuring these future transport systems will be safe during their operation is a mandatory duty. This paper proposes a safety analysis Framework to be applied to the future intelligent Roadway Transport Systems (RTS) paradigm. Based on a combined fast and real time simulation approaches, the Framework allows analyzing, in a broad sense, the impacts of concepts, technologies and procedures on RTS safety and efficiency. A Proof-of-Concept (PoC) is provided, where representative RTS scenarios are modeled, simulated and analyzed using OpenDS, Matlab and Sumo+Veins+OMNet++. The RTS scenario elements (environment, vehicles, roadways, and driver) interact with each other. Autonomous and ‘non-autonomous’ vehicles share the RTS infrastructure, and AV is modeled combining a pair of RTS elements “driver+vehicle”, with the control algorithms embedded in the autonomous driver element to manage AV movement. In this PoC, both the autonomous vehicle behavior and the impact of its embedded autonomous control algorithms over the RTS safety and efficiency are analyzed. Concluding, the Framework is capable to deal with the representative future RTS scenarios, mainly regarding to AV, analyzing the impact of concepts in component level (e.g. AI-based algorithms) over system level properties (e.g. safety) and the relationship among properties at various levels of RTS.