基于虚拟流场的智能车辆全局路径规划方法OA
Global path planning method for intelligent vehicles based on a virtual flow field
水流能够在复杂自然环境中自主探索出合理的流向和可行的路径,借鉴该特性,提出一种基于虚拟流场的自动驾驶全局路径规划方法.首先,基于流体动力学理论建立虚拟道路流场的数学模型,并采用计算流体力学方法对流场状态进行数值求解,在大尺度的复杂道路路网中生成多条从起点到终点的全局流线;同时,通过构建规划路径的综合评价函数,从多条流线中筛选出最优的全局路径;最后,设计GNSS/INS组合定位系统采集真实路网数据,构建全局静态道路地图,在校园场景下验证所提方法.试验结果表明,基于虚拟流场的全局路径规划方法能够根据自动驾驶任务规划并筛选出完整、平滑的全局最优路径,避免陷入环境局部极小点的无解问题.
Inspired by the ability of fluids to naturally identify feasible flow directions and paths in complex environments,this paper proposes a global path planning method for autonomous driving based on a virtual flow field.First,a mathematical model of a virtual road flow field is established using fluid dynamics theory,and computational fluid dynamics(CFD)is applied to numerically solve the flow field,generating multiple streamlines from the start point to the destination within a large-scale road network.A comprehensive path evaluation function is then constructed to select the optimal global path from these candidate streamlines.Finally,a GNSS/INS integrated positioning system is designed to collect real-world road data and construct a global static road map.The proposed method is verified through experiments in a campus scenario.The experimental results show that the virtual flow field-based global path planning method can generate smooth and complete optimal paths according to autonomous driving tasks and effectively avoids failure cases caused by local minima in complex environments.
冀杰;潘登;刘剑;赵立军;张博涵
西南大学 工程技术学院,重庆 400715中国汽车工程研究院股份有限公司,重庆 401122西南大学 工程技术学院,重庆 400715重庆文理学院 智能制造工程学院,重庆 402160西南大学 工程技术学院,重庆 400715
交通工程
自动驾驶路径规划流场流体力学有限元方法
autonomous drivingpath planningflow fieldsfluid dynamicsfinite element method
《重庆大学学报》 2026 (9)
25-34,10
重庆市科学技术局农业农村领域重点研发项目(cstc2021jscx-gksbX0003)重庆市研究生科研创新项目(CYS23207). Supported by Agriculture and Rural Research Project from Chongqing Science and Technology Bureau(cstc2021jscx-gksbX0003),and Chongqing Graduate Research Innovation Project(CYS23207).
评论