Driving Decision-Making Analysis of Car-Following for Autonomous Vehicle Under Complex Urban Environment

Abstract

Car-following decision-making is critical to complete driving mission for autonomous vehicles under complex and dynamic urban environment. The complex and dynamic information have a great influence on the autonomous vehicle's car-following decision. This paper proposes to use the Rough Set theory to abstract the car-following rules to support the decision-making of autonomous vehicles under the complex and dynamic urban environment. Firstly, a virtual urban traffic environment is built by PreScan (a simulation environment for developing advanced driver assistant system). The vehicle dynamics is simulated using 6-DOF dynamic model and the experimental data were acquiring based on MATLAB/Simulink. Secondly, the Rough Set theory is proposed to reduce the influence of weak interdependency data, and extract the driver's decision rules. Finally, the result is that: the subject car will slow down or break when the relative distance between the subject vehicle and the leading vehicle D is less than 7.7m, drive at a constant speed when D is under 10.2-12.7m, and accelerate when D is beyond 12.7m. It shows that the Rough Set can extract the driver's decision rules from the combining multiple sensors data and related information providing a theoretical basis for the car-following decision-making under complex and dynamic environment.