A study on steering angle control method under right/left cornering in an intersection

Abstract

Many articles have reported that the white lines painted on both sides of highways are used by automatically driven vehicles to control their lateral movement when driving. They are detected by stereo cameras or radars. And GPS is used as an additional information for improving its robustness and reliability. However, when we think about an urban area, the stereo cameras and radars cannot recognize any white lines because white lines do not exist at intersection. Therefore, a method for controlling a vehicle without environmental recognition becomes necessary. In our method, the steering angle is controlled by using simple map information comprising nodes and links of roads, which are widely used in navigation systems, and the data of the vehicle's position, which is obtained continuously. In addition, by using two front gaze points to solve the problem associated with one front gaze point, we propose a method that allows a vehicle to steer smoothly when turning at intersection. We performed an experiment using an actual vehicle under the same conditions as those used in the simulation. The results confirmed the validity of our proposed method because the target steering angles at each corner and the driving trajectory of the actual vehicle were almost the same as those obtained using the simulation.