Model and Simulation of Driving Decision-Making Behavior in Curve Movement on ACT-R Cognitive Framework

摘要

The method based on ACT-R cognitive framework involved in bend movement is presented in this paper, which defines perception-action modules, memory modules and core unit of decision-making cycle which includes visual and tactile perception of traffic conditions rules. The paper introduces nine components of state to describe bend movement. Bend movement is divided into normal rational driving behavior and non-rational driving behavior which can be affected by drivers' fatigue, characters, etc. This paper also analyzes the impact of actions and requirements of action generation. Utility function is used to evaluate drivers' actions on the basis of driver's perception expressed by the actions under the new state. The validity of these models is tested by a number of simulations, recommended parameter values are given in this paper to illustrate the effectiveness of decision-making behavior based on ACT-R model.