摘要
基于应用型高校产学研平台,构建智能交通V2X通信感知一体化系统,融合多模态环境感知、低延迟高可靠通信及云—边—端协同计算,实现车路协同下的信息交互与智能决策优化。系统采用C-V2X PC5与5G NR-V2X进行低时延数据传输,结合Kalman Filter+GNN进行多源传感数据融合,提升感知精度。计算层依托NVIDIA Jetson AGX Orin+TensorRT 8.5,优化推理效率,降低25%计算时延。强化学习(RL)与图神经网络(GNN)驱动信号控制与路径规划,交叉口通行时间降低23.8%,交通流波动下降18.5%。实证测试表明,系统在复杂环境下具备高精度、低时延与自适应优化能力,为智能网联交通提供高效、稳定的技术支撑。
Based on the application-oriented university Industry-University-Research platform,an intelligent transportation V2X communication perception integrated system is constructed,which integrates multi-modal environment perception,low-delay and high-reliability communication and cloud-edge-end collaborative computing to realize information interaction and intelligent decision optimization under vehicle-road collaboration.The system uses C-V2X PC5 and 5G NR-V2X for low-latency data transmission,and combines Kalman Filter+GNN for multi-source sensing data fusion to improve the sensing accuracy.The computing layer relies on NVIDIA Jetson Agxorin+Tensor RT 8.5 to optimize the reasoning efficiency and reduce the computing delay by 25%.Reinforcement learning(RL)and graph neural network(GNN)drive signal control and path planning,and the intersection transit time is reduced by 23.8%,and the traffic flow fluctuation is reduced by 18.5%.The empirical test shows that the system has high precision,low delay and adaptive optimization ability in complex environment,which provides efficient and stable technical support for intelligent networked transportation.
作者
张舒
牟凯
ZHANG Shu;MOU Kai(Jilin Institute of Architecture and Technology,130114)
出处
《长江信息通信》
2025年第5期44-46,共3页
Changjiang Information & Communications
基金
吉林省教育科学“十四五”规划2024年度基地专项课题“应用型高校智慧交通专业产学研协同育人实施路径研究”(JD2436)。
关键词
V2X通信
车路协同
云边端计算
智能交通优化
V2X communication
vehicle-road coordination
cloud computing
intelligent transportation optimization
作者简介
张舒(1989-),女,吉林长春人,硕士研究生,讲师,研究方向:智慧交通;牟凯(1994-),女,吉林长春人,硕士研究生,讲师,研究方向:智慧交通。
