摘要
驾驶员注意力不集中是导致汽车追尾事故频发的主要原因,采用自主紧急制动系统可以减少碰撞的数量、降低事故严重程度。针对雷达传感器成本高以及对于静止物体存在“虚目标”等问题,设计了一种基于视觉传感器的自主紧急制动算法。该算法利用一个单目视觉传感器根据图像坐标系中车辆的位置和速度计算碰撞时间(TTC),进而判断是否需要施加制动。将障碍物阴影行坐标根据小孔成像模型映射为两车距离,并利用Kalman滤波器将各行像素的测距误差调整为观测误差矩阵。仿真结果表明:在不同工况下,与传统距离估计算法相比,提出的时变观测误差矩阵Kalman滤波算法能有效降低估计误差,当TTC准确值低于3 s时,TTC的估计值能收敛于准确值,说明所设计的AEB算法有效。
Driver’s inattention is the main cause of frequent rear-ends collisions.Autonomous emergency braking system can reduce the number of collisions and the severity of accidents.Aiming at the high cost of radar sensor and the problem of“virtual target”for stationary objects,an autonomous emergency braking algorithm based on visual sensor is designed.The algorithm uses a monocular vision sensor to calculate the collision time(TTC)according to the position and speed of the vehicle in the image coordinate system,and then determines whether the braking is necessary.The line coordinates of obstacle shadow are mapped to the distance between two vehicles according to the aperture imaging model,and the observation error matrix of each pixel is adjusted by Kalman filter.The simulation results show that the proposed Kalman filtering algorithm for time-varying observation error matrix can effectively reduce the estimation error under different conditions compared with the traditional distance estimation algorithm.When the TTC accuracy is less than 3 seconds,the TTC estimation can effectively converge to the accuracy value,which shows that the designed AEB algorithm is effective.
作者
莫夫
MO Fu(School of Mechanical and Electrical Engineering,Guangdong University of Science and Technology,Dongguan 523083,China)
出处
《重庆理工大学学报（自然科学）》
CAS
北大核心
2019年第12期55-60,共6页
Journal of Chongqing University of Technology：Natural Science
基金
广东省普通高校特色创新类项目(2018KTSCX262)
广东科技学院2017年度院级科研重点项目(GKY-2017KYZD-1)
作者简介
莫夫,男,硕士,副教授,电子工程师,主要从事物联网应用、大数据分析及算法研究,E-mail:dgmofuzi@qq.com。
