Improving interference alignment by precoding methods

Interference alignment is a novel technique which enables interference to align at each receptor, using only half of the signal space, and leaving the other half for the intended signal, independent of the number of users that the channel has. Most of the literature focuses on the technique from the scope of information theory, where the achievable rates are demonstrated. We are interested in the improvement of the performance. In this paper we will work with precoding methods such as singular value decomposition and Tomlinson-Harashima precoding with a bit and power loading strategy. Simulations show that using precoding and bit-power loading methods, the BER can be improved compared to the original Cadambe and Jafar scheme.