Study On Broad-Band Spatial Spectrum Estimation Methods At Subarray Level

Phased array system is an important application field of spatial spectrum estimation techniques. In large phased array systems, the array usually consist of thousands of elements, so the subarray structure is usually employed and only the digital subarray outputs can be obtained.Broad-band spatial spectrum estimation techniques become more and more important because broadband signals are widely used in present Radar systems. Many effective methods of direction finding of the broadband signal have been proposed based on continuing improvement of the narrow-band spatial spectrum estimation techniques. In this case, the present proposed direction finding algorithms based on element level are not suitable any more; we can only but use methods at subarray level.Firstly, the influence of frequency bandwidth to direction finding, namely, the frequency-angle ambiguous problem, and one broad-band spatial spectral estimation method ISSM_SL (Incoherent signal subspace Method at Subarray Level) are studied in this paper. Angle ambiguous exacerbates as the frequence bandwidth extends. ISSM_SL can helpfully lessen the ambiguous but with hevy computation load.Secondly, two broad-band spatial spectrum estimation methods which are CSST_SL (Coherent Signal Subspace Transform at Subarray Level) and WAVES_SL (Weighted Average of Signal at Subarray Level) with high resolution and high precision are studied. WAVES_SL performs much better than CSST_SL because it possesses both advantages of MUSIC and WSF. The disadvantage of the two broad-band superresolution methods is that they need prio-estimation of sources directions which may degrade the performance seriously.To reduce the performance degradation caused by the angle prio-estimation another two broad-band spatial spectrum estimation methods STP_SL (Space and Time Processing at Subarray Level) and STD_SL (Steered Delay at Subarray Level) are then studied. STP_SL method is based on spectral and temporal 2-dimensional samplings and the delay network is adopted at subarray level. STD_SL method is spatial only processing method and the steered time delays are adopted at subarray level. These two methods can eliminate the effect of the variable frequency to broad-band spatial spectral estimation. And they need no angle prio-estimation.Then the weighting networks using analytic method and genetic method adopted at subarray level to improve the performance of direction finding methods are studied in this paper because all the broad-band spatial spectrum estimation methods studied above are based on DSAM which can realize superresolution only in the vicinity of mainbeam.In addition, the influence of I/Q channel error and filter error to ADBF and STAP are studied.We simulate and analyze all the presented algorithms. The simulation results demonstrate the efficiency of the algorithms.