Research On DOA Estimation Algorithm In Smart Antenna

Smart antenna is the key technology in the third generation of mobile communication system. It is also one of the hottest research objects of the fourth generation of mobile communication system. DOA (Direction of arrival) estimation technology is an important part of smart antenna. Through DOA estimation technology, smart antenna is able to estimate the spatial information of users, and selectively receive or transmit multiple signals in the same channel to reduce interference among signals. Therefore, the shortage on frequency resources within the communication system is solved, and the capacity and service quality of the mobile communication system are improved greatly. So whether in separating multi-user signals of uplink or in the selective launch of downlink signals, the measurement of DOA for user signals is the essential premise to realize the directional launch of smart antenna.This thesis focuses on studying DOA estimation algorithm in smart antenna. It summarizes the previous work, and proposes some improved algorithms. Major works are as follows:Firstly, an introduction to research status and development trend of smart antenna technology and DOA estimation algorithms at home and abroad is given. There is also a detailed introduction to the basic structure, classification and basic working principle of smart antennas. Then a summary on DOA estimation in smart antennas is given.Secondly, a simulation research and comparative analyses on some typical DOA algorithms are initiated. For one thing, the simulation experiments for delay addition algorithm and Capon algorithm of minimum variance among traditional DOA estimation algorithms are performed. For another, researches on MUSIC algorithm and ESPRIT algorithm among subspace algorithms are also carried out. The performances of all the DOA estimation algorithms are analysed by means of many simulation experiments made with different focuses and stress.Thirdly, a special problem for DOA estimation-DOA estimation of coherent signals is researched specifically. Based on analysis of the classical spatial smoothing algorithm, the MUSIC algorithm is improved by using Toeplitz martix reconstruction and low-rank approximation. The improved Root-MUSIC algorithm is proposed by defining correlation function and introducing new polynomial. Then, the effectiveness of improved algorithms is proved by simulation experiments and comparative analyses.Finally, the summary on the work described in this thesis and the prospect evaluation of future researches to be done are given.