基于SVM的低空飞行冲突探测改进模型

为保障通航飞行器在低空空域的飞行安全，提出了一种基于支持向量机（SVM）的飞行冲突探测改进模型。首先，建立适应于飞行器的保护区。然后，利用改进型ID3决策树算法将搜索空间降低到局部的方法筛选具有潜在飞行冲突的飞行器，并利用随机森林（RF）选择合适训练集。最后，利用tanh函数优化容易饱和的sigmoid函数对SVM分类结果的概率映射。通过仿真验证和对比分析，结果表明:利用基于密度聚类的DBSACN算法去除异常点，将剔除产生误报和虚报的数据作为训练集优化SVM分类器，改进的飞行冲突探测模型的误报率和虚报率分别降低了0.6%和1.9%，算法执行效率得到提高，而且具有较好的抗干扰能力与稳定性。 

Improved conflict detection model of low-altitude flight based on support vector machine

In order to ensure the flight safety of general aviation aircraft in low-altitude airspace, an improved model of flight conflict detection based on support vector machine (SVM) is proposed. First, according to the physical form and flight status of the aircraft, a protection zone suitable for the general aircraft is established. Then, the improved ID3 decision tree algorithm is used to reduce the search space to a local method to select aircraft with potential flight conflicts, and choose the appropriate training set by random forest (RF) method. Finally, the tanh function is used to optimize the probability mapping of the easily saturated sigmoid function to the SVM classification results. Through simulation verification and contrastive analysis, the results show that the DBSACN algorithm based on density clustering is used to remove outliers, and the data generated by false alarms and missing alarms are removed as the training set to optimize the SVM classifier. Therefore, using improved flight conflict detection model, the false alarms and missing alarms are reduced by 0.6% and 1.6% respectively, and the execution efficiency of the algorithm is improved. The model has better anti-interference ability and stability.