Transceiver optimization for multi-user systems : rate improvement and degrees of freedom characterization
Date of Issue2014
School of Electrical and Electronic Engineering
Positioning and Wireless Technology Centre
This thesis is devoted to optimizing the transceiver strategies in multi-user wireless systems for maximizing the users' achievable rates and the degrees of freedom (DoF). Multi-user wireless systems can be modeled as interference channels. Under the assumption that interference is treated as noise at each receiver, the prevalent transmission strategy in the literature is proper or circularly symmetric complex Gaussian (CSCG) signaling, i.e., the transmitted signal vector from each user is CSCG distributed. In the first part of this thesis, we study the achievable rates of Gaussian interference channels when the more general improper or circularly asymmetric complex Gaussian signaling is applied. For the Gaussian multiple-input multiple-output interference channel (MIMO-IC) with interference treated as noise, we show that the user's achievable rate with improper Gaussian signaling can be expressed as a summation of the rate achievable by the conventional proper Gaussian signaling in terms of the users' transmit covariance matrices, and an additional term, which is a function of both the users' transmit covariance and pseudo-covariance matrices. The additional degree of freedom in the pseudo-covariance matrix, which is conventionally set to be zero for the case of proper Gaussian signaling, provides an opportunity to further improve the achievable rates of Gaussian MIMO-ICs by employing improper Gaussian signaling. To this end, we propose widely linear precoding, which efficiently maps proper information-bearing signals to improper transmitted signals at each transmitter for any given pair of covariance and pseudo-covariance matrices. For the two-user single-input single-output interference channel (SISO-IC) and K-user multiple-input single-output interference channel (MISO-IC), efficient algorithms are proposed to optimize the transmit covariance and pseudo-covariance matrices. By utilizing the separable structure of the achievable rate expression, we demonstrate that improper complex Gaussian signaling can provide strict rate gains over the conventional proper Gaussian signaling. In the second part of this thesis, we study the Gaussian interference channel from the DoF perspective. While most of such studies in the literature have assumed full-rank channel matrices; in certain practical scenarios, the channel matrices may be rank-deficient due to poor scattering and the presence of very few signal paths. We provide an exact DoF characterization for the 3-user MT x MR MIMO-IC with rank-deficient channel matrices, where each transmitter is equipped with MT antennas and each receiver with MR antennas, and the interfering channel matrices from each transmitter to the other two receivers are of ranks D1 and D2, respectively. It is found that when the interfering links are weak in terms of the channel ranks, i.e.,D1+D2 <=min(MT;MR), zero forcing is sufficient to achieve the optimal DoF. On the other hand, whenD1+D2 >min(MT;MR), a combination of zero forcing and interference alignment is required for DoF optimality. We show that the DoF characterization obtained in this thesis unifies several existing results in the literature. Besides interference coordination under the interference channel model, a more aggressive interference mitigation scheme is known as network MIMO, or multi-cell cooperative processing (MCP) in cellular networks. In the last part of this thesis, we propose a modified block diagonalization (BD) scheme for MCP-enabled network, so that the achievable sum-rate in the low-to-medium signal-to-noise ratio (SNR) regime is improved as compared to the original BD scheme. Such a rate improvement is possible since the hard constraint of zero inter-user interference imposed by BD is relaxed; instead, interference leakage is allowed and carefully controlled via a rate-constrained power minimization problem. As such, the effective noise power can be suppressed, and the power enhancement problem associated with the BD scheme is mitigated.
DRNTU::Engineering::Electrical and electronic engineering::Wireless communication systems