On the Performance of FDD Cell-Free Massive MIMO with Compressed Sensing Channel Estimation

Cell-free massive multiple-input multiple-output (CF-mMIMO) provides a consistent service experience for the users that use mobile terminals. In frequency division duplexing (FDD) mode, the pilot overhead for acquiring downlink channel state information is an obstacle to the application of CF -mMIMO. In this paper, we analyze the sparsity of the channel coefficients in CF-mMIMO. By ignoring the small value of entries in the channel matrix, the coefficients matrix of FDD CF-mMIMO downlink channel is approximately sparse and can be acquired via compressed sensing. Thus, a small quantity of coefficients is required to be solved in max-min fair power control. Numerical results show that the proposed scheme outperforms time division duplex (TDD)scheme and the FDD scheme with the orthogonal pilot in terms of energy efficiency.