改进RRT算法在未知三维环境下AUV目标搜索中的应用  

Application of the improved RRT algorithm to AUV target search in an unknown 3D environment

作　　者：李娟[1,2] 张韵 陈涛[1,2] LI Juan;ZHANG Yun;CHEN Tao(Institute of Ocean Installations and Control Technology,Harbin Engineering University,Harbin 150001,China;School of Intel-ligent Science and Engineering,Harbin Engineering University,Harbin 150001,China)

机构地区：[1]哈尔滨工程大学海洋装置与控制技术研究所,黑龙江哈尔滨150001 [2]哈尔滨工程大学智能科学与工程学院,黑龙江哈尔滨150001

出　　处：《智能系统学报》2022年第2期368-375,共8页CAAI Transactions on Intelligent Systems

基　　金：国家自然科学基金项目(51809060,5217110503);黑龙江省自然科学基金项目(E2017014).

摘　　要：针对未知水下环境下的自主水下航行器(autonomous underwater vehicle,AUV)目标搜索问题,传统方法搜索速度慢且以解决二维平面下搜索问题为主,本文提出了一种基于改进RRT(rapid-exploration random tree)的未知三维环境目标搜索算法。在搜索方面,分别建立了包括目标存在概率地图、不确定度地图、区域遍历度地图在内的实时地图并设定其更新规则,根据搜索目标建立决策函数;在局部规划方面,将滚动规划与改进RRT算法相结合,规划出到搜索决策点的路径。二者的结合,实现了AUV在三维空间下在线实时搜索。仿真表明,该算法具有较强的遍历能力,提高了三维空间下目标搜索的速度。To solve the target search problem of autonomous underwater vehicles(AUV)in an unknown underwater en-vironment,traditional methods have slow search speed and mainly solve the search problem in a two-dimensional plane.In this study,an unknown three-dimensional(3D)environment target search algorithm based on the improved rapidly exploring random tree(RRT)algorithm is proposed.In the aspect of search,real-time maps,including target existence probability,uncertainty,and area traversal degree maps,are established,and their update rules are set.Moreover,the de-cision function is established according to the search target.In the aspect of local planning,the path to the search de-cision point is planned by combining rolling planning with the improved RRT algorithm.Their combination realizes the real-time online search of AUV in 3D space.The simulation shows that the algorithm has strong traversal capability and improves the target search speed in 3D space.

关 键 词：自主水下航行器 目标搜索 快速扩展随机树 局部路径规划 水下环境 滚动规划 实时搜索 决策 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置]