匹配输电线路巡检需求的无人机选型研究

Fast Matching Selection Lgorithm of UAV Platform Based on Task Demand of Transmission Line Inspection

近年来,无人机在电力行业输电线路巡检方面得到广泛应用,输电线路无人机自动化巡检技术因其独特的优势受到国内外研究机构及相关部门的广泛关注。但随着巡检任务需求的快速增长,以及市面上所能选择的无人机平台及其载荷类型越来越多,导致快速完成无人机平台选型的难度增加,通常需要耗费大量人力完成筛选匹配工作。因此,根据任务需求,如何快速可靠地完成无人机平台选型是输电线路自动化巡检研究领域需要重点解决的问题之一。为了解决该问题,该文提出了3层结构的基于异构神经网络的快速匹配选型模型。在模型第1层,提出改进型的独热向量算法,完成了巡检任务相关因素的数值化,并输出任务需求矩阵;在模型第2层,提出基于时间递归神经网络(LSTM)神经网络结构的参数指标生成模型,根据任务需求矩阵生成所需的载荷相应参数;在模型第3层,提出了基于决策树网络的匹配模型,根据生成的参数指标确定最终的无人机平台参数和型号。最后,通过真实选型案例数据,验证了该模型算法的有效性。

In recent years, unmanned aerial vehicle （UAV） is widely used in power industry transmission line inspection. Power transmission lines UAV automation inspection technology has been attracted extensive attention from domestic and foreign research institutions because of its unique advantages. However, the rapid growth of inspection task demand and more and more choices for UAV platforms make it difficult to complete the UAV platform selection rapidly, which usually cost a lot of manpower. For solving this problem, this paper proposes a three-phase framework based on various types neural networks for rapid selection. In the first phase, the adjusted one-hot vector algorithm is presented to transform the description of inspection task into task matrix; in the second phase, the relationship between task matrix and load parameters is constructed by using long short term memory （LSTM） neural network; in the last phase, the UAV platform selection is achieved with decision tree network. Finally, the real-world data are used to validate the effectiveness of the proposed model.