基于离散化分布密度的无人机粗糙决策方法

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。

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

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.