Interference Alignment with Partially Coordinated Transmit Precoding

Abstract

In this paper, we introduce an efficient interference alignment (IA) algorithm exploiting partially coordinated transmit precoding to improve the number of concurrent interference-free transmissions, i.e., the multiplexing gain, in multicell downlink. The proposed coordination model is such that each base-station simultaneously transmits to two users and each user is served by two base-stations. First, we show in a K-user system operating at the information theoretic upper bound of degrees of freedom (DOF), the generic IA is proper when $K \leq 3$, whereas the proposed partially coordinated IA is proper when $K \leq 5$. Then, we derive a non-iterative, i.e., one shot, IA algorithm for the proposed scheme when $K \leq 5$. We show that for a given latency, the backhaul data rate requirement of the proposed method grows linearly with K. Monte-Carlo simulation results show that the proposed one-shot algorithm offers higher system throughput than the iterative IA at practical SNR levels.