基于A~*引导域的RRT智能车辆路径规划算法  

Guiding-Area RRT Path Planning Algorithm Based on A~* for Intelligent Vehicle

作　　者：冯来春 梁华为[2] 杜明博 余彪[2] 

机构地区：[1]中国科学技术大学自动化系,合肥230027 [2]中国科学院合肥物质科学研究院应用技术研究所,合肥230027

出　　处：《计算机系统应用》2017年第8期127-133,共7页Computer Systems & Applications

基　　金：国家自然科学基金(61503362;91420104);安徽省自然科学基金(1508085MF133)

摘　　要：为了泛化RRT(快速搜索随机树)算法在智能车辆路径规划领域内的应用,解决该算法搜索效率低、最近邻搜索函数不合理等问题,本文提出了一种基于A~*引导域的RRT路径规划算法.该算法将A~*算法与RRT搜索算法进行有效地结合,利用由A~*算法在低分辨率栅格图中生成的最短路径来构建引导域,以提升RRT算法的采样效率;同时在设计RRT算法的最近邻搜索函数时考虑车辆自身约束,以增强搜索树节点选择的合理性.通过仿真实验和实车测试,对该算法的优越性、有效性和实用性进行了验证.This paper proposes a RRT path planning algorithm based on the guiding-area which is generated with the A^＊ algorithm. This algorithm can benefit the domain from the following aspects： the applications of RRT algorithm to the field of path planning for the intelligent vehicle can be improved significantly. The performance of the traditional RRT algorithm can be enhanced by solving some inherent issues, such as low searching efficiency, irrational nearest neighbour searching functions etc. The novel algorithm combines A^＊ and RRT effectively. Based on low resolution grid map, A^＊ algorithm is applied to construct the guiding area, which is used to improve the sampling efficiency. To enhance the reasonableness of the selection of searching tree node, the vehicle＇s constraints are considered in the design of the nearest neighbour searching function. Finally, the superiority, validity and practicability of the proposed algorithm is verified in simulations and experiments with the real vehicle

关 键 词：RRT 采样策略 A^＊ 智能车辆 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]