Partial-Fixed Interference Alignment in Multiuser Interference Channel

For further boosting the spectral efficiency in interference channels, the multipoint simultaneous transmission is getting more and more popular. While, the multipoint transmission results in the great increase of interference, which is usually suppressed via iterative computation. For effectivelymanaging the interference from multiple transmitters, a partial-fixed interference alignment algorithm is proposed, which greatly decreases the computational complexity and obtains precoding matrices of all the transmitters without iteration. In the proposed algorithm, all the transmitters are divided into two sets, T-1 and T-2. The transmitters in T-1 adopt fixed eigen-beamformingmatrices, with which the interference signals from T-1 construct the fixed interference subspace of each receiver. Then, we design the precoding matrices of the transmitters in T-2 so that the dynamic interference subspace from T-2 lies in the fixed interference subspace from T-1. Via the subspace alignment, the linear homogeneous equations for the precoding matrices of the transmitters in T-2 can be constructed and solved easily to obtain all the precoding matrices. Meanwhile, for improving spectral efficiency, the transmitters in T-1 are selected based on the singular values to increase the sum rate. Simulation results indicate that the proposed algorithm has excellent performance compared with several classical iterative algorithms, which includes the computational complexity and system sum rate.