Research On Smart Antenna And Its Implementation On Software Defined Radio Platform

Smart antenna, superior in increasing antenna gain, suppressing interference, andimproving spectrum efficiency, has become one of the key techniques in3G and4Gcommunications. The main functions of smart antenna are twofold: direction ofarrival (DOA) estimation and beamforming. This thesis concentrates on improvedDOA estimation algorithms as well as blind beamforming algorithms, and theirimplementation on software defined radio by use of GNURadio.Firstly, the theoretical foundation of smart antenna systems is introduced. Secondly,the classical MUSIC algorithm and some representative improved MUSIC algorithmsused in DOA estimation are discussed. Since the ability of these algorithms toestimate the number of signal sources is limited by the number of array elements, thatis, when the number of sources is greater than that of array elements, these algorithmsalways fail, an improved MUSIC algorithm based on the principle of Euler’s formulais proposed. Simulation results show that the proposed algorithm can estimate thenumber of sources2times as many as the classical MUSIC algorithm can do, and itsperformance is still good in low SNR region. Thirdly, in order to achieve effectivetracking of the desired signal, this thesis discusses the typical beamformingalgorithms and proposes an adaptive beamforming algorithm based on improvedMUSIC algorithm and linearly constrained minimum variance (LCMV) algorithm.Comparison with the existing non-blind algorithms verifies superiority of theproposed algorithm. Finally, how to implement a smart antenna system on softwaredefined radio platform is tackled in detail. Some modules on GNU Radio are createdto receive the data streams, and the proposed blind adaptive beamforming algorithmis implemented as well as tested on this platform. Experiment results verify theeffectiveness of the whole system.