基于DSmT证据折扣理论的短波多通道信息融合方法

Shortwave multichannel information fusion method based on DSmT evidence discount theory

针对由于短波信道复杂恶劣、通道信息间冲突突出,经典合并方式对短波多通道信息合并效果不佳问题,提出基于DSmT证据折扣理论的短波多通道信息融合方法。以DSmT证据折扣理论作为融合框架,实现对短波多通道信息融合推理判决。首先,归一化OFDM帧的相对包络标准差,并通过相对包络标准差与折扣因子计算得到折扣权重。然后,通过比特软信息提取函数获得比特概率,以比特概率对基本信任分配函数定量赋值来作为证据。最后,以PCR6融合规则对多通道软信息进行融合推理判决。在多种信道条件、多种调制方式下与经典合并方式的对比仿真实验表明,所提方法可以获得明显的误码率增益和误包增益。

In response to the problem of poor performance of classical fusion methods in combining multiple shortwave channel information due to the complexity and harshness of shortwave channels and prominent conflicts among channel in-formation,a shortwave multi-channel information fusion method based on the DSmT evidence discounting theory was pro-posed.The DSmT evidence discounting theory was employed as the fusion framework to achieve fusion,inference,and de-cision-making of multiple shortwave channel information.Firstly,the relative envelope standard deviation of the OFDM frame was normalized,and the discount weight was calculated through the relative envelope standard deviation and the dis-count factor.Then,the bit soft information extraction function was used to obtain the bit probability,which was quantita-tively assigned to the basic belief assignment function as evidence.Finally,the PCR6 fusion rule was applied to fuse,infer,and make decisions on the soft information from multiple channels.Simulation experiments comparing the proposed method with classical fusion methods under various channel conditions and modulation schemes indicate significant gains in bit error rate and packet error.