【摘要】 针对高速公路占道施工传统监管模式存在的一些痛点，研发了一套深度融合物联网(IoT)、高精度毫米波雷达与人工智能(AI)视觉分析的安全智慧监管系统。该系统构建了“警示-警戒-警报”三级动态主动防护体系，并集成了智能预警终端(安小虹机器人)与全流程电子化协同管理平台，实现了对施工区域的全天候实时监控、车辆闯入风险的毫秒级分级预警以及隐患排查整改的线上闭环管理。通过在广东省多条高速公路高流量路段的专项养护、复杂路桥多点同步作业及高频次日常养护等多种典型场景中的应用实践，该系统能够实现在单日18万车流量的极端条件下，有效触发预警超过5.4万次；将施工审批时间从数小时缩短至约20分钟，隐患整改周期从平均1天压缩至不足1小时；通过优化交通组织，有效降低因施工导致的交通拥堵时长约30%。更多还原

【Abstract】 In response to several pain points in the traditional supervision mode of lane-occupying construction on expressways, a safety intelligent supervision system was developed that deeply integrates the Internet of Things(IoT), high-precision millimeter-wave radar, and artificial intelligence(AI) visual analysis. The system establishes a three-tier dynamic active protection framework of "warning-alert-alarm" and integrates an intelligent warning terminal(Anxiaohong Robot) with a full-process electronic collaborative management platform. This enables round-the-clock real-time monitoring of construction zones, millisecond-level graded warning of vehicle intrusion risks, and closed-loop online management of hazard identification and rectification. Through application in various typical scenarios on multiple high-traffic expressway sections in Guangdong Province-including specialized maintenance, complex road-bridge multi-point simultaneous operations, and high-frequency routine maintenance-the system has demonstrated its effectiveness. Under extreme conditions with a daily traffic volume of up to 180,000 vehicles, the system effectively triggered over 54,000 warnings. It reduced construction approval time from several hours to approximately 20 minutes, shortened the average hazard rectification cycle from one day to less than one hour, and reduced traffic congestion duration caused by construction by about 30% through optimized traffic organization.更多还原