Null-space cognitive precoding for heterogeneous networks

Abstract

Small-cells are considered as an effective solution to increase capacity and offload traffic from, the current macro-cell cellular system. Owing to the difficulty and costs involved in acquiring new spectrum licenses, small-cells are expected to coexist with their respective macro-cells, in the same spectrum. This leads to considerable interference between the two systems. Optimum performance, at the macro-cell, is achieved when the small-cell terminals transmit their information over the null- space of the macro-cell link. However, availability, at the small-cell terminals, of the macro-cell channel null-space information requires full-cooperation and thus a high overhead of information exchange. In this study, the cognitive precoding schemes are designed under a limited inter-system information exchange and the constraint that the performance of the macro-cell link is kept close to the case where no small-cell network does exist. Two techniques are considered: a two-bit quantisation precoded and a dual space-frequency coding precoded approach. It is demonstrated that the first achieves a performance close to the full cooperation approach recently proposed, but with very low information exchange requirements. For the second, it is show that both the systems are able to coexist without any inter-system cooperation and with a performance close to the non-coexistence scenario.