Adaptive Beamforming Algorithm Based On Game Theory In MIMO-OFDM Systems

In order to realize mobile broadband, many novel technologies are applied in the future mobile communication systems, namely the fourth generation wireless mobile communication system (4G), such as MIMO, OFDM, and smart antenna. These technologies are becoming the hot issues. Channel capacity and spectrum efficiency can enhance several times by using MIMO technique without paying extra bandwidth and transmitting power. OFDM technique can overcome frequency selective fading and suppress narrow-band interference. Smart antenna technique, for example adaptive beamforming, improves the gain on the direction of signal arrival, and makes deep null steering beams on the direction of interference signal, thus suppressing interference and improving the quality of communication. Using these techniques, we can overcome multipath fading, suppress interference, enhance spectrum efficiency, and improve quality of communication in terms of space diversity, time diversity and frequency diversity.Adaptive beamforming is the key of the smart antenna. Classical adaptive beamforming algorithms includes:least mean square (LMS), recursive least squares (RLS), sample matrix inversion (SMI), Stochastic Gradient Constant Modulus Algorithm (SGCMA), least square constant modulus algorithm (LSCMA) etc. However, all these algorithms have some deficiencies, such as complex computation, large computing amount, and low convergence.This thesis focuses on the adaptive beamforming based on the game theory. After analyzing mutual interference between users, we induce the game theory into MIMO-OFDM systems, thus modeling the adaptive beamforming for the game theory problem and solving the problem. In this model, the weight vector of transmit power is adjusted by the game results between users, and the adaptive beamforming is realized by using this weight. This novel algorithm takes the cooperative game theory to lift user utilities and improve the global performance of the system. At last, simulations are performed and results demonstrate that the proposed algorithm have advantages, such as fast convergence, low bit error rate, and high quality of service.