MUSIC算法在相干探海雷达中的应用

根据海杂波的频谱特性,运用对称阵列方法中提到的思想对传统的MUltiple SIgnal Classification (MUSIC)算法进行了两个方面的改进。一是对要进行特征值分解的矩阵进行重构,以解决有相干信号存在的问题;二是把算法推广到频域中执行,以满足目标信号的背景噪声是海杂波的特征。数值模拟的结果表明改进后的算法可以有效地进行相干雷达海洋目标回波的高分辨率空间谱分析。

Application of MUSIC Algorithm in Coherent Radar Detecting Sea Surface Targets

The MUltiple SIgnal Classification(MUSIC) algorithm for estimating the arrival angles of multiple narrow-band plane wave signals incident at an array of antennas has become increasingly popular since its inception by Schmidt.The algorithm is based upon a vector space interpretation of the convariance matrix formed from the array element outputs and has exhibited a performance superior to that of other so-called "superresolution" estimators at moderate to low signal-to-noise ratio.However the conventional MUSIC algorithm is based on the assumption that no two signal arrivals are perfectly correlated,a condition referred to as coherent.As a consequence,the performance of the MUSIC algorithm has been observed to degrade severely in a coherent or highly correlated signal environment as encountered in modern radar system widely applying coherent techniques. So the first modification of this article is to reconstruct the eigendecomposition matrix to eliminate correlation of incident signals.On the other hand,the conventional MUSIC algorithm is developed from the data model that the background noise is white noise.But in the problem of detecting the sea surface targets,the background noise is sea clutter which can not be considered as white noise.The second modification is put forward according to this point.The algorithm is extended to the frequence domain since in certain frequence field sea clutter can be regarded as white noise.The result of numerical simulation shows that the modified algorithm can apply in coherent radar system detecting sea surface targets,providing high-resolution space spectral analysis to radar reflection from targets.