基于离散化分布密度的无人机粗糙决策方法  

Method on rough decision-making for UAV based on discrete distribution density

作　　者：董慧芬[1] 王晓丽[1] 高庆吉[1] DONG Huifen;WANG Xiaoli;GAO Qingji(Insititute of Robots, Civil Aviation University of China, Tianjin 300300, China)

机构地区：[1]中国民航大学机器人研究所,天津300300

出　　处：《系统工程与电子技术》2019年第1期105-111,共7页Systems Engineering and Electronics

基　　金：国家自然科学基金(51277011/E070303;U1533126;51377161);中央高校基本科研业务费专项资金(3122018C002);天津市自然科学基金(17JCYBJC18200)资助课题

摘　　要：无人机"精准农业"任务要求无人机能够响应环境信息进行自主决策,从而完成喷绘任务。首先,研究了面向作业要求的空间属性设计规则,根据精准作业条件属性因子的值域分布特点及其样本密度,提出了一种基于高斯分布密度的反正切属性精细离散方法。然后,利用粗糙集属性依赖度的算法求解属性约简,去掉冗余的属性,获取精准作业的粗糙决策规则表。最后,通过试验验证了粗糙决策方法的合理性和准确性。Unmanned aerial vehicle(UAV)“precision agriculture”task requires UAVs to respond to environmental information and make decisions autonomously,so as to complete the task of spraying.Firstly,the design rules of space attributes are studied which meet the task requirements,according to the range distribution characteristic and the sample density of the accurate operating condition attribute factor,a discretization method of precise quantization is proposed which is based on arc-tangent attribute and Gaussian distribution density.Then,the attribute reduction based on the attribute dependence of the rough set is solved and used to remove redundant attributes,and the rough decision rules table is obtained for precise operation.The rationality and accuracy of the rough decision-making method is verified by the experiment.

关 键 词：设计规则 离散分布密度 精细量化 决策生成 无人机粗糙决策 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]