General precoder interference alignment design with achievable higher multiplexing gain and reduced channel extension cases

Abstract

In a K-user single input-single output interference channel, it is proved that the upper bound for multiplexing gain (MG) is K / 2 . Based on interference alignment technique, Cadambe and Jafar introduced a scheme known as Original Interference Alignment (IA). This scheme is improved by Choi–Jafar–Chung (CJC) in which for reduced channel extension, higher MG is achievable. These designs requirean infinite number of channel extension to achieve the upper bound. In this study, as a common generalisation of Original IA and CJC, the authors focus on designing new precoding matrices. With the same conditions as in Original IA and CJC, by changing the values of defined variables (t and s), different cases could be generated so that different MGs with different channel extensions become possible. Broadly speaking, setting appropriate values to the assigned variables, the proposed scheme yields the same results as Original IA and CJC schemes. For all other values of t and s parameters better results than Original IA are obtained but for some special t and s parameters better results than CJC scheme are achieved. Thus, it is possible to simultaneously achieve a higher MG with lower channel extension (i.e. the reduction of computational complexity) in comparison with previously designed schemes.