Research Of Relay Technology Based On Massive MIMO

In recent years, Massive Multiple-Input-Multiple-Output (Massive MIMO) technology has received noticeable attention in the wireless communication society. By deploying unconven-tionally large number of antenna arrays, Massive MIMO systems are capable of mitigating un-desirable randomness (e.g., small-scale fading, uncorrelated inter-user interference), thus pro-viding significant gains in both spectral efficiency and power efficiency. Hence, the relay sys-tem with Massive MIMO deserves further study to improve existing wireless communication system.Firstly, the paper introduces a Massive MIMO relay system, where multiple single-antenna users simultaneously communicate with a base station through a fixed relay station. The paper studies the detail of uplink and downlink of proposed system, where the relay station and base station are equipped with large antenna arrays, and designs linear signal processing method for both uplink and downlink, namely, maximum ratio combining and zero forcing for uplink and maximum ratio transmission and zero forcing for downlink. According to the above studies, the paper derives the closed-form of approximated ergodic spectral efficiency for different signal processing methods.Secondly, the paper discusses the power scaling law for different signal processing methods while the number of relay station and base station antennas grows to infinity. The conclusion shows that the transmit power of each user and relay station can be cut down inverse propor-tionally to the number of relay station antennas and base station antennas respectively in the uplink, while the transmit power of relay station and base station can be cut down inverse pro-portionally to the number of relay station antennas and base station antennas respectively in the downlink, without any loss in performance.Thirdly, the paper introduces the power allocation problem of sum approximated ergodic spectral efficiency maximization for different signal processing methods. For the zero forcing method of uplink, the paper solves the optimization problem by the Lagrange multiplier method and water-filling algorithm directly. For other signal processing methods, the paper introduces the general optimization problem, thus analyzes the optimal solution where the users number equal to 2, while obtains the sub-optimal solution by one-dimensional search with the designed algorithm where the users number larger than 2.