Intellectual simulation model for transport systems management and safety database

Abstract

This paper suggests the approach to the developing on-board driver's aid systems (DAS) with the function of the dynamic assessment of the traffic safety. The key feature of the approach is the combination of the intellectual multi-agent simulation method and the virtual reconstruction of the road scene based on the data from various on-board informational subsystems (mainly from the technical vision subsystem). The data received from the on-board subsystems in real time mode is used for the virtual reconstruction of the current road scene, which is displayed in the three-level multi-agent simulation model developed by the authors. The lower level of the model is of the outmost interest, as it reflects a multitude of functioning intellectual agents. The agents in the model, being in constant cooperation with each other, make individual decisions about their movement regime based on the intellectual analysis of the combined data about the traffic flow, and taking into account their objective functions. The suggested model allows conducting virtual monitoring of various types of moving objects in the road scene, which have gone out of the technical vision field due to some reasons. This, enables better analysis of the road scene and anticipates possible road safety risks. For road safety risk assessment, the model makes a prognosis for the traffic situation development and evaluates the safety indicators in the real-time mode. Data collection on the multitude of the scenarios of simulated traffic situation development is conducted in the form of traffic situation — behavioral decision. The collected data is processed using methods of intellectual data analysis to evaluate the compliance of the chosen behavior with the road safety standards. As a result of the analysis, a database of the situational rules for the behavioral decisions is formed.