基于多边检测的一种车道偏离评价方法

An roadway departure warning algorithm based on multiedge preview

针对车道偏离预警系统LDWS(Lane Departure warning Systems)的车道偏离评价方法,目前通用的评价算法由于计算过程中使用了横向速度恒定和转向半径一定的假设,与车辆实际的速度和方向控制特性有较大的差别,因此评价结果误差也较大。本文基于稳态预瞄基本原理,通过改进隶属函数的方法,提出一种基于多边检测的驾驶安全性评价指标车道偏离评价方法。根据车辆的当前稳态特性预测车辆未来时刻到达的位置,对车辆的行驶轨迹进行稳态预瞄。并通过车辆预期轨迹点与车道的横纵向距离对车道偏离进行评价,大幅度提高了车辆预警精度,减少空预警的发生。

At present the current departure warning algorithm utilizes the hypothesis of landscape orientation rate invariableness and turning radius invariableness in compute process, so there are errors in estimating results. In roadway departure warning process of LDWS （Lane Departure Warning Systems）, a new algorithm is established based on basic principle of steady preview. Through estimating the interval between the vehicles anticipates position and roadway to estimate the current state of vehicle, the precision of roadway departure warning evaluation is much improved.