摘要
针对智能网联汽车在复杂交通环境中存在的多源感知融合不足、协同决策实时性差与系统资源调度效率低等问题,提出一种基于AI Agent与6G深度融合的网联汽车空间智能体协同架构。该架构采用分层设计理念,构建了包含环境交互层、多智能体引擎层、模型调度层、资源管理层与6G通信承载层的五层系统结构,支持“感知-通信-决策-控制”闭环协同。设计了感知、决策、通信与控制四大核心功能模块,并提出智能任务调度、资源协同管理、闭环反馈与容错保障机制,以实现多智能体的高效协同与资源动态优化。在高保真仿真环境中进行测试,结果表明,该系统在感知精度、决策正确率、通信延迟及资源利用率等关键指标上显著优于传统5G-V2X(5G-vechicle to everything)及非协同方案,目标检测准确率最高达98.7%,V2V(vechicle to vechicle)通信延迟低于5 ms,并可支持25辆车协同运行,验证了所提架构在提升智能网联汽车整体性能方面的有效性与先进性。
To address the challenges of insufficient multi-source perception fusion,poor real-time collaborative decision-making,and inefficient resource scheduling in intelligent connected vehicles(ICVs)operating in complex traffic environments,this paper proposes a collaborative architecture of collaborative AI agents integrated with 6G technology.The architecture adopts a hierarchical design philosophy,comprising five functional layers:environmental interaction,multi-agent engine,model scheduling,resource management,and 6G communication support.It facilitates closed-loop coordination of“perception–communication–computation–control”processes.Four core functional modules-perception,decision-making,communication,and control–are designed,along with intelligent task scheduling,collaborative resource management,closed-loop feedback,and fault-tolerant mechanisms,to achieve efficient multi-agent coordination and dynamic resource optimization.Validation in a high-fidelity simulation environment demonstrates that the proposed system significantly outperforms conventional 5G-V2X and non-collaborative approaches in key metrics such as perception accuracy(up to 98.7%in target detection),decision-making correctness,communication latency(V2V latency below 5 ms),and resource utilization.The system supports coordinated operation of 25 vehicles,confirming its effectiveness and advancement in enhancing the overall performance of ICVs.
作者
潘正辉
PAN Zhenghui(Shanghai Great Wall Motor Technology Co.,Ltd.,Shanghai 200335,China)
出处
《德州学院学报》
2025年第6期43-49,共7页
Journal of Dezhou University
关键词
智能网联汽车
6G通信
多智能体协同
协同决策
资源调度
intelligent connected vehicles
6G communication
multi-agent collaboration
collaborative decision-making
resource scheduling
