2-D Music Algorithm DOA Estimation Based On The Array Transformation

As the main content of the signal processing research, spatial spectrum estimationaims at obtaining the exact locations of signal sources. Spatial spectrum represents thepower distribution. The spatial spectrum estimation has a strong ability in spatial resolving.The main research direction is the system which formed by a serial of sensors. Thesesensors received the signal from the sources, by which the parameters of the signal areextracted. The most important parameters are: Spatial parameters, the azimuths and theelevation angles of the signal sources.Establish a spherical coordinates in a three-dimensional space, an exact location of asignal source can be obtained by knowing the elevation angle, the azimuth and the distancereferring to the origin of the axis.1-D spatial spectrum estimation can estimate oneparameter of signal source in one plane. However, when signal sources are in3-Denvironment, the precise location cannot be estimated by1-D estimation. Therefore, theestimation requires2-D DOA estimation. Thus, the2-D DOA estimation can achieve ahigher accuracy than1-D DOA estimation.2-D DOA estimation plays an important role in military and civilian fields. This paperstudies in-depth on the2-D far-field narrowband sources DOA estimation, The paperillustrates a parallelepiped array in order to satisfy the rotational invariance by thetransformation of the array. Then use the smoothing algorithm to achieve the estimation ofspatial spectrum. From the experiment can be seen that the parallelepiped array has a lowerRMSE than the cruciform array in the same SNR. Proposed an improved fast spatialspectrum estimation algorithm for the issue of the tremendous computational load of theMUSIC algorithm. The improved algorithm reduces the computational load effectively byturning the angle domain DOA estimation into transformation domain. the calculation ofthe improved algorithm is only half of the traditional MUSIC algorithm and the accuracy ishigher which can prove that the improved algorithm has a better performance than thetraditional MUSIC algorithm.