基于深度学习的汽车防追尾预警系统设计  

Design of automobile rear end collision prevention warning system based on deep learning

作　　者：叶浩 徐今强 皮雨蒙 左康渝 YE Hao;XU Jinqiang;PI Yumeng;ZUO Kangyu(School of Electronic and Information Engineering,Guangdong Ocean University,Zhanjiang 524088,China;School of Economics and Management,South China Agricultural University,Guangzhou 510642,China)

机构地区：[1]广东海洋大学电子与信息工程学院,广东湛江524088 [2]华南农业大学经济管理学院,广东广州510642

出　　处：《现代电子技术》2024年第14期157-161,共5页Modern Electronics Technique

摘　　要：当前主流的防追尾系统普遍存在识别因素单一、对驾驶员自身预防效果欠佳的问题,为此,设计一种基于深度学习的汽车防追尾预警系统。将YOLOv5剪枝技术、注意力机制、PID优化器等方法融入网络模型的训练中,以优化模型精度并减小模型体积;其次,以距离判断为主,速度、加速度、事故危害性判断为辅来计算车辆的追尾风险,每次预警后通过MQTT协议将数据上传至物联网平台,并在系统结束运行时对驾驶员进行安全分析。系统最终部署在TensorRT环境上进行再次优化。实验结果表明,所设计的汽车防追尾预警系统响应速度快,适应性强,判断风险较为准确。In allusion to the current mainstream rear end collision prevention systems,there is a common problem of single identification factors and insufficient prevention effect on the driver themselves.A deep learning based automotive rear end collision prevention warning system is designed,which can integrate YOLOv5 pruning technology,attention mechanism,PID optimizer and other methods into the training of network models to optimize model accuracy and reduce model volume.The rear end risk of vehicles is calculated mainly based on distance judgment,and supplemented by speed,acceleration,and accident hazard judgment.After each warning,the data is uploaded to the Internet of Things platform by means of the MQTT protocol,and safety analysis is conducted on the driver when the system ends running.The system is ultimately deployed in the TensorRT environment for the further optimization.The experimental results show that the designed car rear end collision prevention warning system has fast response speed,strong adaptability,and accurate risk assessment.

关 键 词：汽车防追尾系统 深度学习 YOLOv5 物联网 追尾判断 风险预警 

分 类 号：TN911.23-34[电子电信—通信与信息系统] TP302.1[电子电信—信息与通信工程]