Interference Alignment With Partial CSI Feedback in MIMO Cellular Networks

• Published in: IEEE transactions on signal processing Vol. 62; no. 8; pp. 2100 - 2110
• Main Authors: Xiongbin Rao, Lau, Vincent K. N.
• Format: Journal Article
• Language: English
• Published: New York, NY IEEE 15.04.2014; Institute of Electrical and Electronics Engineers; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Alignment; Applied sciences; Coding, codes; CSI feedback dimension; Detection, estimation, filtering, equalization, prediction; Exact sciences and technology; Feasibility; Feedback; Filtering; Filtration; IA feasibility condition; Information, signal and communications theory; Interference; interference alignment (IA); Manifolds; Mathematical analysis; Measurement; MIMO; MIMO cellular networks; Minimization; Optimization; partial CSI feedback; Quantization (signal); Sampling, quantization; Signal and communications theory; Signal to noise ratio; Signal, noise; Telecommunications and information theory; Wireless communications; Wireless networks; Parameter estimation; CSI feedback dimension; partial CSI feedback; Wireless telecommunication; Transmitter; Feedback regulation; interference alignment (IA); Cell network; Coding; Signal quantization; Channel estimation; Pretreatment; Cost efficiency analysis; Mobile radiocommunication; MIMO system; MIMO cellular networks; Filtering; Coding circuit; Signalling; First order; Optimal solution; Telecommunication network; IA feasibility condition; Signal processing; Feasibility; Metric; Signal to noise ratio
• ISSN: 1053-587X; 1941-0476
• DOI: 10.1109/TSP.2014.2307283

Interference alignment (IA) is a linear precoding strategy that can achieve optimal capacity scaling at high SNR in interference networks. However, most existing IA designs require full channel state information (CSI) at the transmitters, which could lead to significant CSI signaling overhead. There are two techniques, namely CSI quantization and CSI feedback filtering, that reduce CSI feedback overhead. In this paper, we consider IA processing with CSI feedback filtering in MIMO cellular networks. We introduce a novel metric, namely the feedback dimension, to quantify the first order CSI feedback cost associated with the CSI feedback filtering. The CSI feedback filtering poses several important challenges in IA processing. First, there is a hidden partial CSI knowledge constraint in IA precoder design, which cannot be handled using conventional IA design methodology. Furthermore, existing results on the feasibility conditions of IA cannot be applied due to the partial CSI knowledge. Finally, it is very challenging to find out how much CSI feedback is actually needed to support IA processing. We shall address the above challenges and propose a new IA feasibility condition under partial CSIT knowledge in MIMO cellular networks. Based on this, we consider the CSI feedback dimension minimization subject to the degrees of freedom requirements, and further propose an asymptotically optimal solution and derive closed-form trade-off results between the CSI feedback cost and IA performance in MIMO cellular networks.