考虑复杂地形的越野环境无人车辆路径规划研究  

Study on Path Planning for Off-road Autonomous Vehicles in Complex Terrains

作　　者：聂士达 刘辉[1] 廖志昊 谢雨佳 项昌乐[1] 韩立金[1] 林思豪 NIE Shida;LIU Hui;LIAO Zhihao;XIE Yujia;XIANG Changle;HAN Lijin;LIN Sihao(School of Mechanical Engineering,Beijing Institute of Technology,Beijing 100081)

机构地区：[1]北京理工大学机械与车辆学院,北京100081

出　　处：《机械工程学报》2024年第10期261-272,共12页Journal of Mechanical Engineering

基　　金：国家自然科学基金资助项目(52002212,52394262)。

摘　　要：无人驾驶车辆在越野环境中行驶时,往往面临着复杂的地形和多变的路面。为进行可靠且高效的路径规划,保证车辆安全机动行驶,提出一种考虑复杂地形的越野环境无人车辆路径规划方法。该方法包括全局路径规划与轨迹规划,针对全局路径规划,提出一种基于崎岖地形人工势场的改进Theta^(*)算法,该算法综合考虑坡度、地面类型、地面高程等因素,使车辆尽量远离崎岖地形,通过减少途经路径的坡度和起伏地形,来提高车辆在越野环境中的通行效率、稳定性及安全性。针对局部轨迹规划,提出面向多变行驶场景的自适应概率路线图算法(Adaptive probabilistic roadmap method,APRM),通过设计不同的采样策略来适应车辆在越野环境中行驶场景以及障碍物的变化,从而提高复杂越野环境局部轨迹规划算法构建路径网络图的效率。通过试验验证,改进Theta^(*)算法输出的全局路径平均坡度减少35.63%,地表起伏程度降低33.56%,APRM算法在非结构化道路与开阔场景下进行局部轨迹规划所用时间分别减少79.68%和54.74%。When autonomous vehicles operate in off-road environments,they often face complex terrains and constantly changing road conditions.To realize reliable and efficient path planning and ensure the safe and maneuverable operation of the vehicles,a path planning method for off-road autonomous vehicles that takes into account complex terrains is proposed.The method consists of global path planning and trajectory planning.For global path planning,an improved Theta^(*) algorithm based on rough terrain artificial potential fields is proposed.This algorithm considers factors such as slope,ground type,and elevation to keep the vehicle away from rough terrains.By reducing the slope and undulating terrains in the path,the efficiency,comfort,and safety of the vehicle in off-road environments are enhanced.Regarding local trajectory planning,an adaptive probabilistic roadmap method(APRM)algorithm is presented for handling dynamic driving scenarios.It utilizes different sampling strategies to adapt to the changing off-road driving conditions and obstacles.This enhances the efficiency of constructing the path network for complex off-road environments.Experimental verification shows that the improved Theta^(*) algorithm reduces the average slope of the global path by 35.63%and decreases the surface undulation by 33.56%.The APRM algorithm reduces the time for local trajectory planning in unstructured roads and open terrains by 79.68%and 54.74%,respectively.

关 键 词：轮式车辆 越野环境 自动驾驶 全局路径规划 局部轨迹规划 

分 类 号：U469[机械工程—车辆工程]