Research on Optimized Speed Limit Model to Improve the Cornering Ability of Intelligent Driving

Abstract

This article analyzes the cornering control principle of the intelligent driving system in detail, and analyzes the input factors that affect its cornering performance. The results show that the system has different requirements for the desired cornering speed limit under different curve radius. Especially in the case of small radius with medium and low speed, the vehicle’s own driving feedback parameters and the same change of road curvature have weaker speed limit ability of the system, and there is also a strong hysteresis. Therefore, it is difficult to achieve a better speed limit effect by relying on the existing curve speed limit model.

Based on the above analysis, this article proposes a method of “speed-curve radius” segmentation to establish an optimized curve speed limit model. During the SACC activation process, the system can automatically calculate the lateral acceleration using the curve radius detected by the camera, and extract the minimum target speed using the speed limit model. Subsequently, the ego vehicle calculates the longitudinal deceleration in real time according to the target speed, and sends a deceleration request for automatic deceleration. When the road curve radius is small, only ACC can be activated, and the detected curve radius cannot be used for stable speed limit control. At this time, the system mainly uses the speed limit model to predict the target speed limit value to ensure the deceleration ability of the ACC function in the longitudinal control. In addition, at certain speeds and curves, when the current model speed limit fails to make the corner smoothly, a test can be used to manually calibrate the lateral acceleration in advance, and the target speed limit value can be calculated by Model 3 to enforce the speed limit to ensure smooth cornering.