Optimal degrees of freedom for the K user MIMO interference channel even with delayed CSIT

Certain practical limitations, such as finite CSI feedback delay, make the knowledge of instantaneous perfect CSI impossible and bring the term of delayed CSIT to the forefront. In this paper, it is described how both current and delayed CSI could be exploited by generalizing the space-time interference alignment transmission algorithm (STIA) to the K (K ≥ 3) user M × N interference channel. Specifically, the degrees of freedom (DoF) gain is characterized as a function of the CSI feedback delay γ and the system parameters (K, M, N). Surprisingly, the DoF become optimal, i.e., no loss is achieved with respect to the case of perfect current CSIT, as far as the normalized CSI feedback delay is less than a specific value dependent on the number of users, and the number of transmit antennas equals to N(K - 1). The proposed algorithm reveals unprecedented properties characterizing the performance of the system by providing the limits of the multiplexing gain with relation to the feedback time delay.