The present unprecedented increase in the number of users of wireless systems has urged the development of tructures with higher capacity, reliability and quality of service. To this end, various techniques have been proposed, including the celeberated “Multiple Input, Multiple Output” (MIMO) structure. In Massive MIMO the number of antennas increases and as a result, the dependence of the system performance on factors such as uncorrelated noise and small scale fading diminish, and more degrees of freedom for better control are provided. In the recent years, base stations cooperation has been introduced as a promising solution to improve spectral and energy efficiency in MIMO structures. There are various senarios for base stations cooperation. One of the most challenging arrangements is coordinated beamforming. In this arrangement, only channel state information (CSI) is shared among base stations and there is no need for their symbol synchronization. Also, the data stream for each user is required to be processed at the user’s base station. Since in practical situations, the channel is estimated and consequently, the perfect CSI is not available at base stations, in this thesis the problem of coordinated beam forming with imperfect CSI is addressed. We formulate the problem as a convex optimization problem with the aim of minimizing total downlink transmit power subject to signal to noise plus interference (SINR) constraint for each user, and use the linear minimum mean square error (LMMSE) estimator for channel estimation. Then, an algorithm is proposed for solving the problem for MIMO structure with limited number of antennas using Lagrange duality and proceed to develop an efficient algorithm for solving the problem in massive MIMO structures by obtaining asymptotic equivalents using random matrix theory. The simulations results confirm the effectiveness of the proposed algorithm. Keywords : Massive MIMO, Base Stations Cooperation, Coordinated Beamforming, Linear Minimum Mean Square Error, Lagrange Duality, Random Matrix
