Ship doa bearing estimation method based on complex blind source separation

Abstract

The invention discloses a ship DOA orientation estimation method based on plural blind source separation, comprising the steps of: (1) estimating the quantity of ships in a detectable area based on cross validation; (2) estimating an array response pseudo-inverse matrix based on plural blind source separation; and (3) estimating a target ship DOA orientation. The method can detect a target ship DOA orientation only by receiving signals instead of emitting any signals, while traditional DOA estimation methods belong to the technology of actively detecting targets. The method can estimate the quantity of ships in an antenna array detectable area, while traditional DOA estimation methods require known target numbers; the estimation precision of the invention does not limited to the number of snapshots, while the estimation quality of traditional DOA estimation methods is highly limited by the number of snapshots, so that the method of the invention is simpler and more optimized.

Description

technical field [0001] The invention relates to the field of transportation, in particular to a ship DOA orientation estimation method based on complex blind source separation. Background technique [0002] Problem statement for ship DOA position estimation. Such as figure 1 As shown, in the equidistant linear array, m and d represent the number of array elements and the distance between array elements, respectively, and the receiving array is located in the far-field area of ​​n ships, and m≥n. Suppose the source signal s(t) sent by n target ships = [s 1 (t),s 2 (t),...,s n (t)] T are narrow-band signals that are independent of each other and have zero mean value, and record the angle between the straight line that reaches the first array element and the normal direction of the array as θ i (i=1,2,...,n), this included angle is called the wave arrival azimuth (angle), that is, the DOA azimuth of the ship. If the first array element is regarded as a reference array, t...

AI Technical Summary

Problems solved by technology

The MUSIC method has good robustness to noise, but it requires a sufficient number of snapshots of the received signal, and its estimation accuracy is mutually restricted by the azimuth difference between target sources to be located; ESPRIT technology can be applied to In the case of a large number, it not only has poor noise suppression ability, but also requires a sufficient number of snapshots of the signal like the MUSIC estimation method

The likelihood function of maximum likelihood estimation is a nonlinear multi-mode function, and it is a very difficult and complicated problem to optimize this objective function

Images