基于改进近似极大似然估计的弯道车速预警

超速行驶是道路曲线段事故发生的主要致因,基于智能车辆的提前预警系统是行之有效的安全辅助措施。提出了一种新的弯道车速预警算法,其采用改进的近似极大似然估计拟合曲线半径,改进算法采用最小平方估计取代了传统的平均3点估计。通过模拟实验得到,在不同数量的采样点、噪声和弧长下,改进的估计算法的精度都要高于传统的极大似然估计。根据预估的半径,车辆控制器便能得到相应的安全行驶车速。最后论文根据车辆实际车速和安全车速阈值的差值计算了车辆控制器发出预警信号时车辆与曲线入口之间的距离。

Curve Speed Warning Based on Improved Approximate Maximum Likelihood Estimation

Speeding is the main causes of road accidents on curves,and early warning system based on intelligent vehicle is an effective measure to guarantee road safety.The study presents a new algorithm for curve speed warning,it introduces improved approximate maximum likelihood estimation（MLE） to fit curve radius,the improved algorithm replace the traditional average of 3 points estimate by least squares estimates.Through simulation experiments,with different number of sampling points,noise and arc length,accuracies of the improved algorithm are higher than traditional method of MLE.According to the estimated radius,the vehicle controller can obtain corresponding safe driving speeds.Finally,according to the difference between actual vehicle speed and threshold of security speed,the study calculates the distances between entrance of curves and the sites where the vehicle controller sent an early warning signal.