Path Planning and Decision Making for Autonomous Vehicle in Urban Environment

In this paper, we focus on the need for autonomous vehicles' path planning and decision making in the urban environment. In our modified method, traffic lanes, road edge and geographic information system(GIS) are integrated to generate three structured road models, which can be used to reduce car's lateral positioning error and guarantee car's driving stably without a precise pre-computed map for localization. Based on these models, car's position, attitude and speed are taken into account to find an optimal guided path and a suitable behavior strategy to keep our car obeying traffic rules. Thus, in the whole architecture, the instability of some modules' detection results and low positioning accuracy of GPS receiver is resolved. Applied to our autonomous vehicle-IN2BOT, which has taken part in 'Chinese Intelligent Vehicle Challenge' for many times, this method is proved more effective than our previous method after experiments under various conditions.