目标运动状态识别与关联匹配是智能车辆环境感知中目标跟踪的关键因素,本文中以单线激光雷达为主要传感器开展了车辆前方目标跟踪算法的研究.在最近邻法基础上构造了包含距离、尺寸和反射强度的"关联函数"进行量测值与目标值之间的关联匹配,并基于改进的当前统计模型设计了自适应卡尔曼滤波估计算法来识别目标的运动状态.仿真和试验结果表明,该目标关联匹配与运动状态识别算法能有效地抑制测量过程中引入的量测噪声,并实时、可靠地估计前方障碍物的位置、速度和加速度,实现对目标的跟踪.%Target motion state recognition and correlated matching are the key factors of target tracking in the environment perception of intelligent vehicles. In this paper, a novel vehicle front target tracking algorithm is studied with single-line laser radar as main sensor. On the basis of nearest neighbor method,the"correlation func-tion"including the distance,size and reflection intensity is constructed to conduct the correlation matching between measurement and target. Then an adaptive Kalman filtering estimation algorithm is designed based on improved cur-rent statistical model to identify target motion state. The results of simulations and tests indicate that the target corre-lation matching and motion state recognition algorithms proposed can effectively suppress the measurement noise and real-time and reliably estimate the position,velocity and acceleration of front obstacles to achieve target tracking.
展开▼
