Semi-parameter number estimation method for coherent and incoherent mixed signals

Abstract

The invention discloses a semi-parameter number estimation method for coherent and incoherent mixed signals. The method comprises the following steps: firstly, determining ranks of a cross covariance matrix and a combined matrix, and estimating the number of incoherent signals and the number of coherent signal groups; secondly, estimating elevations of the incoherent signals by utilizing a linear operator and the cross covariance matrix of two uniform linear arrays, and further calculating an oblique projection operator according to the estimated elevations of the incoherent signals; finally, constructing a telescopic matrix sequence only containing coherent signal information by utilizing the oblique projection operator, and determining the number of signals in each coherent signal group according to the rank sequence of vector product matrixes of telescopic matrixes. According to the method, the number of the incoherent signals and the number of the signals in each coherent signal group are estimated respectively, the influence of superposition noise is reduced, and complicated characteristic decomposition is avoided. Under the condition of small snapshot number and / or low signal to noise ratio, the method is lower in calculation complexity and higher in result accuracy for the number of the incoherent and coherent mixed signals with similar detection distance.

Description

technical field [0001] The invention belongs to the field of array signal processing and can be used for estimating the number of coherent and non-coherent mixed signals. Background technique [0002] The performance of many high-resolution DOA estimation methods depends on prior knowledge about the number of signals. For example, the MUSIC and ESPRIT methods assume that the number of signals is known or estimated in advance. Therefore, the number of signals is estimated in the field of array signal processing. is crucial. In addition, in practical applications, due to the influence of multipath propagation, the incident signal is usually a mixed signal of coherent and non-coherent (independent or partially correlated), so the estimation of the number of signals includes the estimation of the number of non-coherent signals, the number of coherent groups And estimate the number of signals in each coherent signal group. [0003] The existing signal number estimation methods ...

AI Technical Summary

Problems solved by technology

In order to estimate the number of coherent signals, some parameter estimation methods combining signal number detection and direction of arrival estimation have been proposed. However, these methods are usually based on maximum likelihood direction estimation and include a multi-dimensional nonlinear optimization process, and the calculation is very complicated.

If the array is a uniform linear array, some non-parametric estimation methods can also be used to estimate the number of coherent signals. The most famous methods are to use spectral smoothing or forward and backward preprocessing methods to reduce the coherence of the signal, or to use the translation of the uniform linear array The rank of the covariance matrix of the coherent signal is restored without deformation. Although these methods avoid the direction of arrival estimation process, the computational complexity is relatively low, but the performance is poor and because the effective array aperture is reduced, the maximum number of signals that can be processed also less

In addition, although the parametric and non-parametric signal number estimation methods mentioned above can estimate the total number of mixed signals, they cannot estimate the number of non-coherent signals and signals in each coherent signal group separately, even if the smooth rank profile (SRP) method can do this, but its performance is usually poor at low SNR and / or small snapshots

Images