Channel Estimation and Data Detection for Multicell Massive MIMO Systems in Correlated Channels

In this study, multicell multiuser multiple-input multiple-output systems are considered with very large number of antennas at base station. We consider a physical channel model comprised finite number of angle of arrivals, and propose a channel estimation scheme based on the eigenvalue decomposition of the covariance matrix of the received signal vector. To improve the channel estimation performance under strong intercell interference, a joint channel estimation and optimal multiuser detection scheme is proposed. Furthermore, an optimal linear receiver is developed for the uplink by taking the channel estimation error into account, and derive a closed-form expression for the sum achievable rate of the optimal receiver. Moreover, using the channel reciprocity of a time-division duplex system, we present how to form an optimal precoding vector for the downlink, and how to compute the sum achievable rate in the downlink. Numerical results are provided to verify the analysis and evaluate the performance of the channel estimation and data detection.