基于贝叶斯决策风险的元胞蚁群算法路径规划  

A Cellular Ant Colony Algorithm for Path Planning Using Bayesian Decision Risk

作　　者：王秀芬 杨盛毅[2] 田茂祥 胡馨丹 WANG Xiu-fen;YANG Sheng-yi;TIAN Mao-xiang;HU Xin-dan(College of Mechanical Electronic Engineering,Guizhou Minzu University,Guiyang Guangzhou 550025,China;Key Laboratory of Pattern Recognition and Intelligent Systems of Guizhou Province,Guizhou Minzu University,Guiyang Guangzhou 550025,China)

机构地区：[1]贵州民族大学机械电子工程学院,贵州贵阳550025 [2]贵州民族大学贵州省模式识别与智能系统重点实验室,贵州贵阳550025

出　　处：《淮阴师范学院学报（自然科学版）》2020年第4期298-305,共8页Journal of Huaiyin Teachers College;Natural Science Edition

基　　金：贵州省科学技术基金项目(黔科合基础[2017]1088)。

摘　　要：为解决无人飞行器(UAV)采用传统蚁群算法路径规划收敛速度慢、容易陷入局部最优的问题,提出了一种基于贝叶斯决策风险的元胞蚁群算法.将环境地图栅格化,采用尖角优化策略从元胞邻居模型中初步确定备选节点的可行区域.将由蚁群算法获取的信息素抽样概率作为先验概率、将目标节点的启发信息融入条件概率、将障碍物的斥力信息融入风险函数.综合多项因素来评测各备选节点的优越性.依据平均风险的大小进行决策分析,其中扇形预测区域和尖角优化策略的引入使得路径搜索更符合无人飞行器的飞行特征,并且加快了路径搜索速度.仿真实验表明,新算法具有更好的全局搜索能力且达到了更好的避障效果.In order to solve the shortcomings of slow convergence and easy to fall into local optimum when using traditional ant colony algorithm for UAV path planning,a cellular ant colony algorithm based on Bayesian decision risk is proposed.Firstly,the environment map is rasterized,and the feasible candidate nodes are preliminarily determined from the cell neighbor model by the sharp angle optimization strategy.Secondly,the pheromone sampling probability obtained by the ant colony algorithm is taken as the prior probability,the heuristic information of the target node is incorporated into the conditional probability and the repulsion information of obstacles is integrated into the risk function.Thus,the superiority of each candidate node is evaluated by combining multiple factors.Finally,decisions are made based on the average risk.The introduction of the sector prediction area and the sharp angle optimization strategy makes the path planning more in line with the flight characteristics of the UAV and speeds up the speed of path search.Simulation experiments show that the new algorithm has better global search ability and achieves better obstacle avoidance effect.

关 键 词：贝叶斯决策 元胞蚁群算法 扇形预测区域 平均风险 路径规划 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TP24[自动化与计算机技术—控制科学与工程]