基于K-Medoids提取信道状态特征的无人机探测方法  

Drone detection method based on K-Medoids to extract channel state characteristics

作　　者：宋玲玉 潘鹏[1] 刘天乐 SONG Lingyu;PAN Peng;LIU Tianle(School of Communication Engineering,Hangzhou Dianzi University,Hangzhou 310018,China)

机构地区：[1]杭州电子科技大学通信工程学院,浙江杭州310018

出　　处：《电信科学》2025年第1期75-87,共13页Telecommunications Science

基　　金：国家自然科学基金资助项目(No.62301200);浙江省自然科学基金资助项目(No.LQ22F010004);浙江省属高校基础研究基金资助项目(No.GK219909299001-016)。

摘　　要：对低空目标的有效管控是推动低空经济发展的关键。城市环境中强杂波和建筑物遮挡等因素使得传统雷达探测手段难以实现对低速无人机的有效监测。基于此,提出了一种无人机探测的新思路,即通过识别信道状态特征的变化来判断无人机是否出现在指定区域。该方法的核心在于利用城市中已广泛部署的移动基站等外辐射源,基于K-Medoids聚类算法捕捉无人机出现后对原有多径信道路径数量的影响,从而实现对无人机的感知。该方法不需要构建精确的参考信号,也不需要利用多普勒体制抑制强杂波。仿真结果表明,所提方法在1 km~2范围内能实现80%以上的检测概率,且随着范围缩小,检测概率能达到90%左右,因此能够在城市场景下有效探测低空慢速无人机。Effective management of low-altitude targets is key to promot the development of the low-altitude economy.In urban environments,strong clutter and building occlusion make it difficult for traditional radar detection methods to effectively monitor low-speed drones.Based on this,a new approach of drone detection was proposed,which involved identifying changes in channel state characteristics to determine whether a drone was presented in a specified area.The core of this method lied in utilizing the already widely deployed mobile base stations and other external radiation sources in cities,capturing the impact of drone presence on the number of multipath channel paths by using the K-Medoids clustering algorithm,to achieve drone perception.This method did not require the construction of an accurate reference signal nor the use of Doppler systems to suppress strong clutter.Simulation results show that the proposed method can achieve detection probabilities of over 80%within a range of 1 square kilometer,and the detection probability can reach about 90%as the range decreases.Therefore,it is capable of effectively detecting lowaltitude,slow-moving drones in urban scenarios.

关 键 词：无人机 信道状态信息 外辐射源 K-Medoids算法 

分 类 号：TN929.5[电子电信—通信与信息系统]