利用改进DBSCAN算法的管制雷达目标标定方法

为探究基于眼动数据的管制雷达目标标定方法,利用雷达管制模拟机、眼动仪等搭建实验平台,召集8名管制学员参加模拟航空管制实验并收集眼动数据.使用具有噪声的基于密度的聚类算法(DBSACN,density-based spatial clustering of applications with noise)处理注视点数据过程中,主观输入的参数会导致无法很好完成聚类,笔者提出基于K-最邻近算法和变密度阈值设定法,从自适应选取邻域值和变密度阈值设计两方面对DBSCAN算法进行改进,并实现了算法的自适应运行.对改进过程中采用拟合分布密度函数极值点结合放大系数确定邻域值的方法进行验证,发现对不同雷达目标误差仅为8.6％和10％,表明改进方法具有一定的适用性.通过比对不同航空器目标兴趣区的提取结果,发现提出的管制雷达目标标定方法具有一定的准确性和普适性.

Control radar target calibration method using improved density-based spatial clustering of applications with noise (DBSCAN) algorithm

In order to explore the target calibration method of air traffic control radar based on eye movement data, the radar control simulator and eye tracker and other devices were used to build an experimental platform. Eight controller students were recruited to participate in the simulation control experiment and eye movement data were collected. Based on the collected eye movement data, the radar target calibration was performed by using the DBSCAN algorithm. However, it is found that the subjective input parameters will lead to the inability to complete the clustering well. Therefore, we propose to improve the DBSCAN algorithm from the two aspects of adaptive selection of neighborhood values and variable density threshold design based on the K-nearest neighbor algorithm and the variable density threshold setting method. And the adaptive operation of the algorithm is realized. The method of determining the neighborhood value by using the fitting distribution density function extreme point and the amplification factor in the improvement process was verified. And the error of other radar targets is only 8.6% and 10%, indicating that the improved processing method has certain applicability. By comparing the extraction results of different aircraft target areas of interest, it is found that the proposed method of calibrating radar targets has certain accuracy and universality.