Abstract: In expressway toll region, the traditional method of path planning and intelligent decision-making cannot realize safe driving of autonomous driving vehicle due to randomness of surrounding vehicles behavior. To solve the problem, a structured control model, named as decision tree, is proposed, which controls the vehicle speed to make it drives through the toll region along the predefined path safely. Firstly, the scene parameters of the expressway toll region, the method for predefining the path between the toll window and of the ramp, and the vehicle model are given. On this basis, a structured control model of the decision tree is proposed to control the speed to realize safe driving of the vehicle. Meanwhile, the pure tracking method is adopted for navigating the vehicle driving along the predefined path. Finally, random road environment is constructed by computer simulation for verifying the safety of the decision tree control model, the continuity of vehicle speed and feasibility of acceleration and deceleration. The results show that autonomous driving vehicle can pass through the expressway toll region safely with vehicle speed controlled by the decision tree control model under the condition that surrounding vehicles obey traffic rules. In addition, the vehicle speed, acceleration and deceleration comply with the traffic rules, and are within the executable range

Key words: path planning, decision tree, autonomous driving, intelligent decision-making, human-like driving