一种基于混合交通流不确定性的无人集卡感知增强方法研究

打开文本图片集

中图分类号：U463.69 文献标识码：A 文章编号：1003-8639（2026）03-0058-03

【Abstract】Aiming at the challnges posed by the behavioral uncertainties of dynamic obstacles in mixed trafic flow environments tothe operational efficiencyand safetyof unmanned container trucks，this paper proposes a path planning method that integrates vehicle-road cooperative perception and distributed decision-making.The system architecture incorporates onboardmulti-sourcesensors （forenvironmental perception，high-precisionpositioning，and V2X communication）androadside perceptionunits，constructingareal-time traffcenvironment situation through data fusion. Thecoreof themethod liesfirstinbuildingamulti-modaldata-fused intentdecision tree modeltoidentifyand predictthe movement intentions of surounding free-moving targets such asvehicles andpedestrians.Secondly，adistributeddeadlock detectionandreservation framework isdesignedtoeffectivelyidentifyandresolvepotentialpathconflicts through intervehiclecoordination.Onthis basis，the methodcombines global task-level routeplanning with localreal-timetrajectory planning： generating global routes acording to upper-level operational instructions，and dynamicallyoptimizing local trajectories basedonreal-time environmental perception，intent prediction，and deadlock resolution.Simulation and real-scenario testsshowthatthismethodcansignificantly improve thetraficefficiencyanddriving safetyofunmanned container trucks in complex mixed trafic flows.

【Key words】autonomous terminal truck； path planning; driving intention recognition

0 引言

当前，面对全球港口持续增长的吞吐量需求以及日益严峻的驾驶员短缺、人力成本和作业安全挑战，采用无人驾驶技术的集装箱卡车（无人集卡）正成为提升港口自动化水平、优化运营效率、降低综合成本并保障运输安全的关键解决方案之一，并在世界多个先进港口加速从技术测试迈向商业化试运营与规模化部署。（剩余3906字）