基于峰均功率比的盖氏圆信源数估计方法

西安工程大学电子信息学院,陕西西安710600

信号检测; 信源数估计; 特征向量; 峰均功率比; 盖氏圆半径; 不等强多目标

Gerschgorin disk estimator method for source number estimation based on peak-to-average power ratio
JIAO Yameng, WU Yue, and CUI Lin

College of Electronic Information, Xi'an University of Engineering, Xi'an710600, Shaanxi Province, P.R.China

signal detection; source number estimation; eigenvector; peak-to-average power ratio; Gerschgorin disk radius; multi-target with unequal intensity

DOI: 10.3724/SP.J.1249.2021.01085

备注

针对低信噪比下盖氏圆估计(Gerschgorin disk estimator, GDE)方法难以有效估计信号源数目的 情况,通过将峰均功率比(peak-to-average power ratio, PAR)与盖氏圆半径相结合,提出基于峰均功率比的盖氏圆估计(Gerschgorin disk estimator based on peak-to-average power ratio, GDE-PAR)方法.从阵元接收数据中提取出特征向量,再用特征向量加权接收数据,得到峰均功率比用于修正对应的盖氏圆半径,用GDE方法判定信号源数目.在不同信噪比条件下对修正前后的盖氏圆半径进行Matlab仿真,并使用蒙特卡罗方法对不同信噪比、等强双目标与不等强双目标和等强3目标与不等强3目标情况下,Akaike信息论准则(Akaike information criterion, AIC)、最小描述长度(minimum description length, MDL)准则、PAR方法和GDE-PAR方法进行信号源估计.结果表明,在低信噪比及不等强多目标条件下,GDE-PAR方法成功检测到信源的概率均比AIC、MDL准则和PAR方法大.

Gerschgorin disk estimator(GDE)method is difficult to detect the number of signal sources effectively under low signal-to-noise ratio(SNR). In view of this situation, we propose GDE method based on peak-to-average power ratio(GDE-PAR)by combining the peak-to-average power ratio(PAR)with the Gerschgorin disk radius. We extract the eigenvectors from the received data and use them as the weights of the data. The PAR is used to modify the corresponding Gerschgorin disk radius, and the number of signal sources is determined by GDE method. This paper adopts Matlab simulation on the Gerschgorin disk radius before and after correction under different SNR conditions. In the case of double target of equal intensity and double target of unequal intensity, three target of equal intensity and three target unequal intensity using Monte Carlo method, we use Akaike information theory criterion(AIC), minimum description length(MDL)criterion, PAR method and GDE-PAR method to estimate the number of signal sources respectively. The results show that under the conditions of low SNR and multi-target of unequal intensity, the probability of successful detection of source by GDE-PAR method is higher than that of AIC, MDL criterion and PAR method.

·