Study On Method Of Active And Safe Virtual Driving Behavior Based On Information Fusion

Intelligent Transport System (ITS) provides an efficient method of modern traffic and transport management to improve traffic environment and efficiency with intelligent information technology. The transport system is synthesized by people, vehicles, roads and environment, and people are initiative and play a key role as the intermediate link between complex environment and vehicles. The study of driver's behavior process in the environment of multi-source information is the key of the research on ITS.This paper studies the traffic environment and drivers'behavior process in the virtual driving experiment system, and focuses on the drivers' behaviors and influencing factors in three stages of information perception, judgment & decision and operation behavior, and proposes a mechanism to evaluate drivers'safety based on the priority of safety factors. A behavior decision module is established for the drivers in the multi-source environment based on the information fusion theory and method. This model consists of safety evaluation module, early-warning module and behavior reasoning module. Active driving behavior evaluation model which is key function of behavior reasoning module is completed using BP neural network technology. The experiment shows the mapping between multi-resource fusion information and driving behaviors is established successfully using behavior reasoning module, and it is feasible to introduce BP neural network algorithm into active driving behavior model. Finally, the virtual driving experiment system is completed, including scene design, function requirement analysis, interaction scheme and user interface design.