基于最小临近点迹和航迹关联的多源目标融合方法

针对智能车辆多传感器的目标融合问题，提出了一种改进的基于欧氏距离与余弦相似度的点迹和航迹数据关联的车用多传感器目标跟踪融合算法。该方法需获取由毫米波雷达系统和Mobileye视觉系统检测到的目标物数据列表，并对两个传感系统检测到的目标物数据进行匹配关联；然后对目标物进行匹配跟踪，更新目标物的生命周期状态；最后对上述两个传感系统输出的目标物的数据进行融合。该算法能够融合视觉系统和雷达系统两个传感系统的优点，以达到精确感知环境信息的目的，从而解决单一传感器难以满足感知系统精度及可靠性需求的问题。

Multi-Source Target Fusion Method Based on the Nearest Positioning and Trajectory Data Association

Aiming at tracking and fusion with multiple sensors for intelligent vehicles, the paper proposes an improved vehicle multi-sensor target fusion algorithm based on positioning and trajectory data association using the Euclidean distance and cosine similarity. Initially the lists of target data detected by the millimeter wave radar system and the Mobileye vision system are obtained, and the targets detected by the two sensing systems are matched and associated. Then the tracked targets are matched and the lifecycle states of the target are updated. Finally the target output from the above two sensing systems are fused. The algorithm in this paper integrates the visual system and the radar system to achieve accurate perception of the environment, superior to the single sensing system in terms of accuracy and reliability