Research On DOA Estimation Algorithm Of Smart Antenna In TD-SCDMA System

The resource of wireless spectrum becomes insufficient with the increasing mobile customers. As one of the key technology of the TD-SCAMA communication system, smart antenna is more attractive nowadays. The technology of smart antenna leads the research direction to the field of air separation, which hence provides an effective solution to ease the inadequate of the spectrum, to enhance the communication capacity and service quality and to minimize the signal interferences and multipath effect. The DOA, one of the key processes of smart antenna, is of great significance of the directional emission of downlink signal, directly making the precise prediction of the mobile-users DOA as the primary premise of the smart antenna.Among the existing DOA estimation algorithms, the multiple signal classification (MUSIC) algorithm and Estimation of Signal Parameters via Rotational Invariance Technique (ESPRIT) algorithm are classic Subspace algorithms, such algorithms can accurately detect the signal DOA under ideal conditions with high estimation accuracy and resolution. However, when the SNR is low or the incoming signal is coherent source, the performance of Subspace algorithm will be greatly affected, and even failure. In allusion to the limitations of these algorithms, three improved algorithms are presented in this paper, which are MUSIC algorithm based on eigenvalue criteria and eigenvalue difference criteria, and the ESPRIT algorithm for detection of coherent sources.The algorithms in this paper perform decoherence using reconstruction of Toeplitz matrix. The improved MUSIC algorithm performs decoherence to output covariance matrix of the antenna array, while the improved ESPRIT algorithm performs decoherence to eigenvectors of the covariance matrix. With detailed theoretical derivation and a large number of simulation experiments, the characteristics of the improved algorithms, the classical subspace algorithms and the existing algorithms are analyzed. Simulation results show the strengths and effectiveness of the algorithm provided in the detection of coherent sources.