Study On No-Blind Adaptive Beamforming Technology Based On Second Order Cone Programming

Adaptive beamforming technology is one of key technology for the smart antenna,non-blind adaptive beamforming technology has been widely used because it’s goodperformance in theory. However, due to the non-idealized actual array antenna, one main jobof smart antenna research is using lesser operation and simpler structure to achieve adaptivebeamforming, therefore, an even better adaptive beam forming algorithm should be obtainedto solve the large amount of computation and structural complexity, to this end, the non-blindadaptive beamforming technology is studied baced on second order cone programming(SOCP) method in the thesis, to achieve the purpose of excellent beamforming performance,convergence speed and small operation costs.1. In analysis of the commonly used non-blind adaptive beam forming technique: theminimum mean square (LMS) algorithm, the recursive least squares (RLS) algorithm and thesample matrix inversion (SMI) algorithm. The simulation comparison results showed: LMShas simple structure, but the convergence the speed is subject to many restrictions; RLS hasgood convergence, but computation is enormous; SMI has fast processing speed, but there iscumbersome matrix inversion processing, and has memory overhead promblem.2. The convex optimization problem model and SOCP algorithm standard form areanalyzed, and the least squares planning problem is found to be a special class of convexoptimization problem that can be solved by transforming into an SOCP problem. Thus,solving algorithm used SOCP is simple and reliable.3. Sampling matrix inversion algorithm based the SOCP (SOCP-SMI) is proposed,namely: SMI problem model is solved by converting into SOCP problem, this can effectivelyavoid the matrix inversion process and greatly reduce the complexity in the SMI algorithm.Through simulation experiments, the side lobe suppression, the weight vector iteration and theaspects of computational complexity are be comparative analyzed, results show that th theSOCP-SMI algorithm can effectively suppress sidelobes and improve the performance ofadaptive beamforming; and furthermore, it has faster convergence rate, lower computationalcomplexity, so it is more effective than the classic SMI algorithm.