CAJ | 학술논문

利用信息融合技术研究多系统调制识别问题,提出一种加权表决融合算法.该算法首先定义系统对识别结果的贡献度,用来描述多个系统的联合对识别性能的提升,并将其作为分配权重的依据.然后将这种多系统识别问题建模为合作对策模型,利用合作对策的Shapley值来决定各系统的权重.最后通过所提出的加权表决算法,对多个系统的调制方式判决结果进行决策层融合.计算机仿真表明该算法有助于提高调制识别的性能.
이용신식융합기술연구다계통조제식별문제,제출일충가권표결융합산법.해산법수선정의계통대식별결과적공헌도,용래묘술다개계통적연합대식별성능적제승,병장기작위분배권중적의거.연후장저충다계통식별문제건모위합작대책모형,이용합작대책적Shapley치래결정각계통적권중.최후통과소제출적가권표결산법,대다개계통적조제방식판결결과진행결책층융합.계산궤방진표명해산법유조우제고조제식별적성능.
Information fusion technique is utilized for modulation classification and a weighted voting algorithm is presented. Firstly,this algorithm defines a contribution to classification results to describe performance improvement of multisystems and depends on it to assign weights for each system. And then the modulation classification of multisystems is modeled as cooperative games and the weights are decided by Shapley value of cooperative games. Finally, by using the weighted voting algorithm presented, decision fusion results of modulation classification is processed. Computer simulation demonstrates an effective improvement in recognition performance via this algorithm.