基于RVM的无人机轨迹异常检测研究OA
Research on UAV Trajectory Anomaly Detection Based on RVM
在无人机异常检测领域,针对传统无人机(Unmanned Aerial Vehicle,UAV)轨迹异常检测算法因欠缺参数选择导致欠拟合、检测准确率下降问题,文中分析了高维飞行数据间的相关性,提出了一种基于RVM(Relevance Vector Machine)的无人机轨迹异常检测算法TADMR(Research on UAV Trajectory Anomaly Detection Based on RVM).通过设计无人机轨迹特征并结合随机森林算法进行特征分析,提取特征组成轨迹特征向量,构建RVM异常检测模型.实验结果表明,TADMR算法的异常检测准确率为94.16%,召回率为96.21%,与传统的无人机轨迹异常检测算法相比更有效.
In the field of UAV(Unmanned Aerial Vehicle)anomaly detection,aiming at the problems of un-derfitting and decreased detection accuracy caused by the lack of parameter selection in traditional UAV trajectory anomaly detection algorithms,this study analyzes the correlations among high-dimensional flight data and proposes a UAV trajectory anomaly detection algorithm TADMR(Research on UAV Trajectory Anomaly Detection Based on RVM)based on RVM(Relevance Vector Machine).By designing the trajectory features of unmanned aerial vehi-cles and conducting feature analysis in combination with the random forest algorithm,the features are extracted to form trajectory feature vectors,and an RVM anomaly detection model is constructed.The experimental results show that the anomaly detection accuracy rate of the TADMR algorithm is 94.16%and the recall rate is 96.21%,which is more effective compared with the traditional unmanned aerial vehicle trajectory anomaly detection algorithm.
田崇旭;付雄;邓松;张冲标
南京邮电大学 计算机学院,江苏 南京 210023南京邮电大学 计算机学院,江苏 南京 210023南京邮电大学 碳中和先进技术研究院,江苏 南京 210023国网嘉善供电公司,浙江 嘉善 314199
信息技术与安全科学
无人机模式识别RVM特征提取异常检测随机森林
UAVpattern recognitionRVMfeature extractionanomaly detectionrandom forest
《电子科技》 2026 (9)
13-18,6
国家自然科学基金(51977113)国网浙江省电力有限公司科技项目(5211JX240001)National Natural Science Foundation of China(51977113)Science and Technology Project of State Grid Zhejiang Electric Power Co.,Ltd.(5211JX240001)
评论