Precoder Design For Physical-Layer Multicasting In Wireless Networks

With the continuously development of wireless communications, the number of wireless users and wireless service are growing rapidly. However, the available spectrum resources in the wireless communication systems are quite limited. To meet the demands on wireless service, it is very important to develop advanced communication techniques with high spectral efficiency. As a result, physical layer multicasting is proposed. By utilizing the property of broadcasting in wireless channels, physical layer multicasting can send common messages to multiple nodes simultaneously. Compared to conventional point-to-point transmission, the multicasting scheme saves the transmission resources, therefore significantly improves the system throughput and also spectral efficiency. Due to these favorable characteristics, physical layer multicasting has received significant attentions in recent years.In the multiple-input multiple-output (MIMO) communication systems, if both the transmitter and the receivers have perfect knowledge of the channel state information (CSI), then precoding at the transmitter can improve the performance of physical layer multicasting. Therefore, this dissertation focuses on the precoding techniques in the multicast scenario. Firstly, the beamforming design for MIMO multicasting is investigated to maximize the minimal signal-to-noise ratio (SNR) of all users, where only one data stream is transmitted. Furthermore, when all the users are equipped with multiple antennas, dual-stream multicasting and multiple-stream multicasting are considered and efficient precoding methods are developed. The main works and contributions of this dissertation are listed in the following:1) For the multicast scenario where one data stream is sent, efficient beamforming approaches are developed for the cases of two transmit (Tx) antennas and more than two transmit antennas at the transmitter.When the transmitter is equipped with two Tx antennas, by deriving the feasible set of the SNR vector of all users, the original complex-valued optimization problem is transformed into a real-valued optimization problem. Based on this result, the global optimal beamforming vector can be found by exhausting a group of hypothesis tests. In order to reduce the complexity of exhausting, a prune and search algorithm (PASA) is developed to find the global optimal beamformer. In addition, the computational complexity of PASA is carefully analyzed.When the transmitter is equipped with more than two Tx antennas, an iterative two-dimensional optimization (I2DO) algorithm is proposed, which iteratively transforms the original problem of beamformer design into a sequence of two-antenna subproblems. Hence PASA can be used to improve the beamformer at each iteration of I2DO, and near optimal beamforming vector is obtained finally.The simulation results show that the proposed beamforming methods have superior performance to most of the existing beamforming techniques, moreover, their computational complexity is also much lower.2) For the multicast scenario where two common data streams are sent, two efficient precoding approaches are proposed, respectively, for the zero-forcing decision feedback equalizer (ZF-DFE) receivers and the minimum mean-squared-error decision feedback equalizer (MMSE-DFE) receivers.When the ZF-DFE receivers are employed at all users, the unitary precoder is designed to maximize the minimal SNR of all subchannels. After the feasible set of the SNR vector for both data streams and also the relationship between them are derived, the original matrix optimization problem is converted into a real vector optimization problem. Then a gradient-based iterative algorithm is developed to calculate the unitary precoder efficiently.When the MMSE-DFE receivers are employed at all users, the precoder is designed on the basis of the criterion of maximizing the minimal signal-to-interfere-and-noise ratio (SINR) of subchannels. Since the precoder is constructed by a transformation matrix and a rotation matrix, the design of precoder includes two stages. First, the transformation matrix is calculated under the criterion of maximizing the throughput of the multicast channel. Once the transformation matrix is obtained, the gradient-based iterative algorithm can be applied to obtain the rotation matrix. In particular, for the special case of two users, a constructive method is proposed to directly calculate the rotation matrix.Both theoretical analysis and simulations show that the proposed approaches outperform the existing ones in both the achievable rate and symbol error rate.3) For the multicast scenario where multiple common data streams are sent, two precoding approaches are developed, respectively, for the flat fading multicast channel and the frequency-selective fading multicast channel, respectively.For the flat fading multicast channel, unitary precoder design to maximize the minimal SNR of all data streams is considered. With the special structure of this problem, the Givens rotations are employed to update the precoder matrix iteratively. Whereas in each iteration, the original problem of multiple-stream multicasting is transformed into a dual-stream one. Therefore, the gradient-based iterative algorithm can be used and finally the nearly optimum precoder is obtained.When the multicast channels are frequency-selective channels, to maximize the system throughput and also minimize the group delay, the minimum-phase finite impulse response (FIR) precoder design is investigated. This problem is tackled in two steps. Firstly, based on the criterion of maximizing the throughput, two efficient algorithms for the nonminimum-phase FIR precoder design are proposed, respectively, from perspectives of frequency domain and time domain. In the second step, based on the theory of spectral factorization, the nonminimum-phase FIR precoder is transformed into the corresponding minimum-phase FIR precoder by a classic iterative algorithm without affecting the throughput.Numerical results indicate that the symbol error rate of the proposed unitary precoder is much lower than that of the existing methods, besides, the achievable rate of the proposed FIR precoder has remarkable improvement over the existing schemes and the group delay introduced by the FIR precoder is also minimized.