可达集约束下的自主车辆路径规划势场蚁群算法研究  

Potential Field ant Colony Algorithm for Autonomous Vehicle Path Planning Under the Constraints of Reachable Set

作　　者：杨海洋 胡辛 郑福银 吕俊波 Yang Haiyang;Hu Xin;Zheng Fuyin;LüJunbo(Department of Transportation,Xi’an Jiaotong Engineering Institute,Xi’an 710300,China)

机构地区：[1]西安交通工程学院交通运输学院,西安710300

出　　处：《黑龙江科学》2025年第4期84-87,共4页Heilongjiang Science

基　　金：陕西省大学生创新创业训练计划项目“复杂环境下车辆自主决策安全性验证”(2024DC84)。

摘　　要：为了解决当前路径规划技术无法涵盖汽车全部未知状况导致降低其安全性能的问题,提出一种以后向可到达集合为限制条件的自动驾驶最佳路线优化方案,将后向可到达集合的变化范畴设为势场蚂蚁算法的制约因素,在多个车队行驶环境中应用此特性,通过观察后向可到达集合各安全子区域的信息素密度差异,发现距离风险区更近的地方信息素密度较低这一特点,构建出一种适合自动驾驶的最优路线模型。实验结果显示,此策略不但提升了传统的势场蚂蚁算法的安全保障能力,还能计算出自动驾驶车辆行进过程中的安全位置可能达到的范围,并对未来一定时间内自动驾驶车辆的安全情况做出预估,说明在复杂环境中可利用势场蚁群算法和混合系统来确定可达集,实现路径规划。To solve the problem that the current path planning technology cannot cover all unknown situations of the vehicle and thus reduce its safety performance,the study proposes an optimal route optimization scheme for autonomous driving with the posterior reachable set as a constraint condition.Specifically,we set the scope of change of the posterior reachable set as a constraint factor for the potential field ant colony algorithm and apply this characteristic in a multi-vehicle driving environment by observing the information density difference of the safe subregions in the posterior reachable set and discovering that the information density is lower in places closer to the risk area.We construct a suitable optimal route model for autonomous driving,while ensuring that the algorithm has a high confidence value in its safety reliability.The experimental results show that this strategy not only improves the safety assurance ability of the traditional potential field ant colony algorithm,but also calculates the possible range of safe positions for the autonomous driving vehicle during its travel process and makes a forecast of the vehicle’s safety situation within a certain period of the future.In complex environment,path planning is a key problem in intelligent transportation,and one solution is to use potential field ant colony algorithms and hybrid systems to determine the reachable set.

关 键 词：智能交通 路径规划 可达集 复杂环境 势场蚁群算法 混合系统 

分 类 号：X913.3[环境科学与工程—安全科学] U471.15[机械工程—车辆工程]