多感知融合的后装汽车防碰撞系统研究

Research on Rear-Mounted Vehicle Collision Prevention System Based on Multi-Sensing Fusion

摘要: 设计了一种多感知融合的外加装汽车防碰撞装置。针对当前存量车辆主动安全配置的普遍缺失，导致复杂交通场景下碰撞风险居高不下。本文提出一种基于多传感器融合的后装防碰撞系统，通过有限度的车辆改装实现主动安全功能升级。系统采用毫米波雷达与视觉传感器的异构数据互补策略，在Haar-like特征检测框架中引入卡尔曼滤波算法，并构建碰撞时间和安全车距的复合型安全模型。硬件层面设计非侵入式改装方案：毫米波雷达嵌入前保险杠非承力区域，视觉模块利用后视镜基座实现无损安装，控制单元通过车载诊断接口(OBD-II)获取车辆状态信息，并经由CAN总线协议触发分级制动响应。研究成果为汽车后市场安全升级提供了兼具工程可行性与经济性的技术路径。

Abstract: An external anti-collision device with multi-sensing fusion was designed. In view of the general lack of active safety configuration of current vehicles, the collision risk remains high in complex traffic scenarios. A rear-mounted anti-collision system based on multi-sensor fusion is proposed in this paper. Active safety functions can be upgraded through limited vehicle modification. The system adopts the heterogeneous data complementary strategy of millimeter wave radar and vision sensor, introduces Kalman filter algorithm into the Haar-like feature detection framework, and constructs the composite safety model of collision time and safety distance. Hardware level design of non-invasive modification scheme: millimeter wave radar embedded in the front bumper non-bearing area, vision module using the rearview mirror base to achieve non-destructive installation, control unit through the on-board diagnostic interface (OBD-II) to obtain vehicle status information, and through the CAN bus protocol to trigger a staged braking response. The research results provide a technical path of engineering feasibility and economy for the safety upgrade of automotive aftermarket.