Interference alignment is a promising tool to reduce the impact of interference in the communication network. However, it requires instantaneous channel state information at transmitters (CSIT), an extremely challenging condition to be met, which makes the technology impractical.

In this thesis, I propose several transmission schemes under different scenarios and investigate the achievable Degrees of Freedom (DoF), which captures the number of independent data transmitted per time slot, for different multi-user channels with a Multiple-Input Multiple-Output (MIMO) relay in the network with no CSIT.

Firstly, I start with the 2-user interference channel. I explore the characteristics of the interference channel and see how relay helps to increase the DoF. I apply the technology of i...[ Read more ]

Interference alignment is a promising tool to reduce the impact of interference in the communication network. However, it requires instantaneous channel state information at transmitters (CSIT), an extremely challenging condition to be met, which makes the technology impractical.

In this thesis, I propose several transmission schemes under different scenarios and investigate the achievable Degrees of Freedom (DoF), which captures the number of independent data transmitted per time slot, for different multi-user channels with a Multiple-Input Multiple-Output (MIMO) relay in the network with no CSIT.

Firstly, I start with the 2-user interference channel. I explore the characteristics of the interference channel and see how relay helps to increase the DoF. I apply the technology of interference cancellation to obtain an achievable sum-DoF, which is a lower bound for the sum-DoF of the network. My result also shows that unlike channels with CSIT where adding relay does not increase the DoF performance, when channel state information (CSI) is not available at transmitters, having a relay can improve the DoF performance of interference channels. Then I generalize the case to k-user channels and use retrospective interference alignment to show how a transmission scheme with proper beamforming construction can achieve good DoF results in general M × N interference channels, where M and N are the number of antennas at each transmitter and receiver respectively.

Since the interference channel is a special case of X channels, I extend the results in the interference channels to find suitable transmission strategies for the X channels. I first consider the symmetric case and assume the relay has sufficient antennas so that it is able to decode all the transmitted signals in one time slot. My result matched the conclusion of the interference channel case, which shows that even without CSIT, my transmission scheme can achieve the optimal sum-DoF as long as the relay has sufficient antennas. I also study the minimum number of antennas required at the relay to achieve the optimal DoF. Moreover, the results are generalized to cases with an arbitrary number of antennas at the transmitter, the receiver, and the relay. As X channel is the most general case and includes the MIMO broadcast channel as special cases, I also compare the DoF results of the X channels with those of the MIMO broadcast channels.

Finally, I switch the transmission from data speech to the video content, which can be cached in advance in the network. In the multi-user MIMO networks with caching, the channel topology will be transferred systematically by placing content in the network node. Thus the system latency, seen as the reciprocal of DoF, can be improved by the transmission scheme as well. A tradeoff between storage and latency in a cache-aided multi-user network has been made by applying the DoF results in the general non-caching multi-user channels.

In this thesis, I propose a general scheme to analyze the DoF for MIMO multi-user channels, including interference channel, X channel, and broadcast channel, also caching network included. By adding a MIMO relay, I make the interference alignment more practical to be implemented. An information theoretical approach is applied to align the interference. Without CSIT, I do not require the feedback of channel estimation to the transmitters. Not only the optimal DoF, when instantaneous CSIT is available, can be achieved by my scheme, promising results for different cases are also derived which are better than existing works. It gives some insights on making interference alignment possible in reality.