直升机林区植保作业飞行进入回避区识别方法

Identifying Helicopter Flight Entering the H - V Diagram in Plant Protection Operations

摘要:

直升机在林区植保作业中具有响应速度快、农药利用率高、防治效果明显等特点，是林区病虫害防治的重要载体。同时，避免进入回避区是直升机在林区作业起飞和着陆阶段的关键，自动识别直升机进入回避区成为一项重要任务。本研究基于飞行参数，运用支持向量机 (Supportvectormachine,SVM) 算法提出了一种适用于林区植保作业直升机飞行进入回避区的快速识别方法，对其低空飞行安全保障和飞行质效评价具有重要的工程价值。通过在直升机高度 - 速度包线内外侧提取数据点作为训练集和测试集，采用交叉验证方法对核函数的参数选择进行优化，从而构建一套基于支持向量机 (SVM) 的直升机进入回避区预测模型，对比分析了多项式核函数与径向基核函数的预测性能差异，阐明了核函数实现高维空间超平面识别直升机回避区内外数据的机理，并应用该模型识别了直升机在回避区内外侧的高度 - 速度组合飞行数据。实验结果显示，尽管多项式核函数及径向基核函数模型在测试集的预测准确率均达到 0.894, 但径向基核函数模型对飞行数据的预测精度为 100%, 优于多项式核函数的预测精度 (97.3%), 表明径向基核函数模型在直升机飞行回避区识别中的泛化能力更佳。

Abstract:

As one of the important vehicles, helicopters play a key role in the forest-protection and pest control task. Compared with unmanned vehicles, helicopter can fly faster, carry more payloads with longer endurance. Avoiding entering the H -V diagram is a key issue for a helicopter's take-off and landing phase, and a significant task is to automatically identify entering the area. A method was developed for helicopters entering the H - V diagram based on support vector machine (SVM) theory, which had significant value for helicopters' safety management and flight evaluation. By selecting some data of a helicopter's H -V diagram as the training and testing groups, and the cross-validation algorithm was used to optimize kermel function's parameters, a prediction model for H - V diagram was developed based on SVM. Both the poly and RBF kernel functions vere adopted for comparing the test results, and also the flight data (height-velocity) around the H -V diagram were identified based on the prediction model. The calculation showed that although the same accuraey (0. 894) was obtained by using the poly and RBF kernel models, the RBF model's predietion accuraey got to 100%,beter than poly kermel model (97.3%), which again showed that the RBF kernel model had enhanced generalization ability. In the future work, the high-speed H - V curve's identification should be emphasized so as to enhance the safety for helicopters in plant protection operations.