基于异类传感器多模型融合方法的研究  

Research on multiple model fusion algorithm with heterogeneous sensors

作　　者：钮俊清[1] 刘军伟[1] 任清安[1] 

机构地区：[1]中国电子科技集团公司第38研究所,合肥230088

出　　处：《空军预警学院学报》2015年第3期194-197,共4页Journal of Air Force Early Warning Academy

摘　　要：低空/超低空的突防对国土防空提出了严峻的挑战.针对融合问题中多源、不同维度和异步等特征问题,通过多个传感器有机地结合,利用多源异类传感器进行数据融合,提出了基于熵的交互式多模型的异类融合方法,并设计了两种异类传感器集中式的数据融合策略.最后通过仿真实验对比分析了两种方法.仿真结果表明采用序贯滤波的融合方法具有较好的融合性能.Low altitude/extreme low altitude surveillance challenges seriously our homeland air-defense. In view of the characteristic problems of fusion including the multi-source, different dimensions, asynchronous and etc., this paper presents a heterogeneous fusion method with interactive multiple model based on the entropy, by combining dynamically the multiple sensors and using the multi-source dissimilar sensor to make data fusion,designs the strategy of data fusion with two types of dissimilar sensor centralized next, and finally, analyzes those two methods through simulation experiments. Simulation results show that using the fusion method with sequential filtering is of better fusion performance.

关 键 词：异类融合 交互式多模型 序贯滤波 

分 类 号：TN957[电子电信—信号与信息处理]