Capacity of Multi-User Inter-Cell Interference Alignment with Measured Channels

It has been shown that in a multi-user multi-cell network, the Multi-User Inter-Cell Interference Alignment (MUICIA) based transmit precoding outperforms the non-alignment based state of the art precoding schemes. The theory of MUICIA is based on the assumption that the partial and stale inter-cell interference (ICI) information is still useful. This ICI is used by the transmitter to design he precoding such that the ICI and the multi-user interference (MUI) are aligned at the receiver. In this paper we provide the proof of this concept with the help of measured channels. By using a measurement test-bed, we demonstrate that in slow time variant channels, the partial ICI contains useful information. Our outcomes are based on two different indoor scenarios for a multi-user multi-cell setup with users which are equipped with multiple antennas. We also evaluate and analyse other multi-user mimo precoding schemes. We evaluate the impact of training overhead on the performance of the precoding schemes in order to realize the feasibility of precodings with nominal overhead. The results show that the concept of MUICIA is valid and it provides higher gains in a multi-cell setup than other non-alignment based precodings with high as well as nominal training overhead.