基于BP神经网络的空中目标航迹预测模型  

Aircraft Target Track Prediction Model Based on BP Neural Network

作　　者：钱夔[1] 周颖[1] 杨柳静[1] 谢荣平[1] 何锡点[1] 

机构地区：[1]中国电子科技集团公司第二十八研究所,南京210007

出　　处：《指挥信息系统与技术》2017年第3期54-58,共5页Command Information System and Technology

摘　　要：针对热点区域的目标航迹预测问题,提出了一种基于反向传播(BP)神经网络的空中目标航迹预测模型。首先,采用基于轮廓系数的自适应K-means聚类算法,将目标群航迹数据自适应聚类,从而提取出特定目标的活动区域变化规律;然后,利用BP神经网络对目标群航迹进行训练学习,建立航迹预测模型,实现目标飞行航迹的提前预测;最后,通过试验结果表明该模型能够有效提取目标群航迹规律并预测目标航迹,具有较强鲁棒性。Aimed at the problem of target track prediction in hotspot area, an aircraft target track prediction model based on back propagation （BP） neural network is proposed. Firstly, an adap-tive K-means clustering algorithm based on silhouette coefficient is adopted. The target group track data are adaptively clustered, thus the change rule of the activity area of the special target is extracted. Then, the target group tracks are trained and learned by BP neural network, and the track prediction model is established to realize the advance prediction of the target flight track. Finally, experimental results show that the model can effectively extract the target group track rule and predict the target track with stronger robustness.

关 键 词：航迹预测 轮廓系数 自适应K-means BP神经网络 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TP183[自动化与计算机技术—控制科学与工程]