Abstract:

The traditional oneside interference alignment (IA) algorithms cannot effectively reserve the desired signals in the interference-free subspace over multiple-input multiple-output (MIMO) interference channels, which causes the degradation of the system capacity. To deal with this, a new IA design concept with auxiliary intermediate variables is put forward, then a method based on adaptive composite cost function (ACCF) is proposed to design the precoder matrices. In the method, the adaptive weighted difference between the powers of the leakage interference and useful signals is firstly defined as the cost function. Then the interference suppression matrices are transformed into the intermediate variables by an auxiliary function in order to obtain the adaptive composite function. Finally, the composite cost function is optimized to obtain the precoder with the modified gradient descent method on the Grassmann manifold. Simulation results verify the effectiveness of the ACCF-based IA method over MIMO interference channels.