面向车联网的城市交通目标级协同感知系统研究  

Research on object-level cooperative perception systems for urban traffic in vehicular networks

作　　者：薛拯 刘畅 韩国军[1] XUE Zheng;LIU Chang;HAN Guojun(School of Information Engineering,Guangdong University of Technology,Guangzhou 510006,China)

机构地区：[1]广东工业大学信息工程学院,广东广州510006

出　　处：《应用科技》2025年第1期98-105,共8页Applied Science and Technology

基　　金：广东省基础与应用基础研究基金广州市联合基金(2022A1515110602);广东省自然科学基金项目面上项目(2023A1515012189);广东省“珠江人才计划”引进创新创业团队项目(2021ZT09X044).

摘　　要：随着车联网技术的快速发展,其在降低单车智能成本、提升道路交通安全及效率方面展现出显著优势。针对自动驾驶中的超视距、盲区和遮挡等感知长尾难题,基于车联网通信设备和自动驾驶计算单元开发了一套具备良好兼容性和可拓展性的目标级协同感知系统,配套的软件环境方便部署检测模型和后期融合算法。通过对传输链路时延测试和感知结果可视化分析,系统能够有效整合感知结果,且能在现有检测模型基础上进行低成本升级,增强自动驾驶感知能力和范围。该系统推动协同感知技术从理论研究向实际应用的转变,具备向实际道路交通系统扩展的潜力。With the rapid development of the Internet of Vehicles(IoV),significant benefits have been observed in reducing the cost of single-vehicle intelligence,enhancing road safety,and improving traffic efficiency.To address the long-tail perception challenges in autonomous driving,such as beyond-line-of-sight scenarios,blind spots,and occlusions,a novel object-level cooperative perception system has been developed based on vehicular network communication devices and autonomous driving computing units,offering excellent compatibility and scalability.Its comprehensive software environment supports the deployment of detection models and the integration of subsequent fusion algorithms.The system's efficacy is demonstrated through latency testing of the transmission link and detailed visualization analysis of perception results.It facilitates the integration of perception data and allows for cost-effective upgrades using existing detection models,thereby extending the perception capabilities and operational range of autonomous vehicles.This advancement underscores the system's potential to bridge the gap between theoretical research and practical application,setting the stage for its integration into real-world traffic systems.

关 键 词：车联网 协同感知 自动驾驶 单车智能 目标级融合 目标检测 检测模型 智能交通 

分 类 号：TN92[电子电信—通信与信息系统]