Introducing a novel safety assessment method through the example of a reduced complexity binary integer autonomous transport model

Abstract

Our previous work focused on the creation of a binary integer model framework aiming at network level traffic optimization through the control of individual vehicles. High computational demand arising from time and space discretization has been identified as the main limitation of the concept. To deal with this, new methods are presented in this paper in order to reduce the complexity of the optimization process with a particular emphasis on system safety. In the first step the most relevant hazards of the system were identified and they were used as the basics of the further development process. The effects are quantified implying a significantly reduced computational demand, without threatening the feasibility of the results. Considering the fundamental requirement of ensuring safe traffic, novel methods are introduced to determine the safety level of the results provided by our model and the described hazard types. The safety indicators defined here cover factors related to the crossing movements, the average speed and average change in speed of vehicles, investigating them also at the network level. The presented methods have significant potential related to the design of real-time, safety-focused, and effective transport management processes of connected and automated mobility systems.