Interference alignment in partially coordinated multipoint receivers

Interference alignment (IA) is a promising technique to achieve more than one degree of freedom in wireless interference networks. It has been shown that in a k-user multiple-input multiple-output (MIMO) interference channel where each N-antenna transmitter communicates with its intended M-antenna receiver by sending d independent data streams, the aggregate number of transmit and receive antennas should satisfy M + N ≥ d(K + 1). This impedes the practical implementation of IA since for large K or d, the required number of antennas for each transceiver pair grows excessively. On the other hand, coordinated multipoint transceivers are inherent parts of future wireless networks since they provide significant performance improvement in communication quality of interference-limited networks. While the performance of coordinated transmitters has been widely investigated in the literature, in this paper and by relying on coordinated receivers, we propose IA algorithms which simultaneously satisfy the following properties: 1) unlike the common assumption of full information exchange among nodes and owing to the partial coordination, averagely half of the total data and cross channel state information is needed to be exchanged among receive nodes, 2) the feasibility condition of the proposed IA algorithms can now be expressed as M + N ≥ d(K + 4)/2 which further implies that for asymptotically large K and to achieve the same number of degrees of freedom, the required number of antennas for proposed methods can be half of that of regular IA techniques, 3) even with this reduced number of antennas, the achievable sum rate of proposed schemes is still comparable to that of conventional IA methods.