Wideband Coherent Source DOA Estimation Based On Music Algorithm

Array signal processing is a very important topic in the field of science, as its toppriority, the spatial spectrum estimation has a wide range of applications in many fields.The research purpose of spatial spectrum estimation is to obtain the positions of signalsources. That is positioning which is essential to study in military surveillance radar andweapons guidance, civil smart antennas, resource exploration and emergency rescue.According to the bandwidth of the sources and the correlation between differentsources to be processed, spatial spectrum estimation can be subdivided for narrowbandand wideband signal or coherent and incoherent. Narrowband incoherent signal spatialspectrum estimation techniques (such as MUSIC,ESPRIT,ML) have been very maturewhile broadband coherent signals estimation is more difficult than narrowbandincoherent signals’, so the methods for coherent wideband signal spatial spectrumestimation are few. However, in reality, broadband coherent signals are not only morecommon but also containing more useful information, so it is essential to solve it. For theshortcomings of original algorithms, the article puts forward a new decorrelationprocessing algorithm and an improved algorithm based on CSM algorithm to overcometheir weaknesses.We often have to pre-process and remove the correlation of datum received bysensor array before processing coherent signals and adopt some incoherent signalsestimation algorithms to estimate. Though decorrelation processing methods are so many,each has its own flaws like spatial smoothing algorithm. It is excellent but still has someflaws such as too much loss of array aperture and low utilization rates of elements. Forthe flaws of spatial smoothing algorithm, the article proposes a new decorrelationprocessing algorithm which recycles every element and makes the most use of theinformation received by each array elementThere are two wideband signal processing methods: One is ISM which simplydecomposes a wideband signal into several narrowband signals and uses ripe narrowbandsignal estimation algorithms to estimate; the other one is CSM which constructs afocusing matrix to remove correlation between different signal sources and uses ISM toestimate. The former method can’t process coherent signals; the latter method is susceptible to signal interference and has too much calculation. Based on CSM algorithm,the article puts forward an improved method which uses toeplitz matrix to reconstruct anew focusing matrix to avoid the interference caused by signal fluctuations, save moretime to calculate and enhance real-time positioning greatly.Above all, the article proposes two algorithms for decorrelation processing andwideband coherent signal spatial spectrum estimation. By a large number of simulations,the effectiveness of the above algorithms will be testified.