Cooperative Techniques for Interference Management in Wireless Networks

In the last few years, wireless devices have evolved to unimaginable heights. Current forecasts suggest that, in the near future, every device that may take advantage of a wireless connection will have one. In addition, there is a gradual migration to smart devices and high-speed connections, and, as a consequence, the overall mobile traffic is expected to experience a tremendous growth in the next years. The multiuser interference will hence become the main limiting factor and the most critical point to address. As instrumental to efficiently manage interference between different systems, this thesis provides a thorough study on cooperative techniques. That is, users share information and exploit it to improve the overall performance. Since multiuser cooperation represents a very broad term, we will focus on algorithm design and transceiver optimization for three cooperative scenarios that capture some of the main features and practical issues: the interference channel (IC the underlay cognitive radio (UCR) model and the two-way relay channel (TWRC). In the IC, K transmitter-receiver pairs exploit the channel state information (CSI) to jointly design their transmit strategies. For this scenario, we provide two main contributions following the lines of interference alignment (IA), whereby linear precoding is applied to eliminate the inter-user interference. First, we consider an ideal model for a multiple-input multiple-output (MIMO) IC that uses several channel extensions (e.g., different subcarriers). In this setting, we design an efficient algorithm to compute IA precoders, which, as opposed to the state-of-the-art algorithms, it guarantees the dimensionality of the signal subspace. In second place, we shift to a more realistic model for the MIMO IC with channel extensions. Specifically, we consider orthogonal frequency-division multiplexing (OFDM) transmissions, where the application of existing IA algorithms require an additional level of cooperation: time synchronization. To avoid such demand, we apply the precoders and decoders at the sample level in the time domain, which allows the users to transmit asynchronously. We propose two different algorithms that are evaluated by means of simulations and real measurements, where their effectiveness is revealed. Then, we move to the UCR model, where the so-called secondary users (SUs) are allowed to coexist with the primary user (PU) as long as the latter achieves a prescribed data rate requirement. To this end, interference constraints are imposed to the SUs, hence significantly reducing the cooperation overhead with respect to the IC scenario. We provide contributions to two different PU settings: single-antenna and multiple-antenna point-to-point links. In the former case, we consider an interference power constraint imposed to the secondary network. We first analyze whether a single-antenna SU can benefit from following an improper signaling scheme, and contribute with several analytical results and a closed-form expression to determine when there is a payoff in terms of achievable rate. Second, we consider several multiantenna SUs and study how to efficiently assign an interference power constraint to each of them, so that their cooperation needs can be reduced with respect to a global or aggregate interference constraint. We provide a novel solution based on the statistics of random projections that requires local CSI. In the multi-antenna PU scenario, not only the interference power is relevant, but also its spatial distribution. Our first contribution for this setting is a closed-form expression for the maximum tolerable interference power given a PU rate requirement, which is obtained by adopting a worst-case assumption on the spatial signature of the interference covariance matrix. Motivated by this observation, we then propose a spatial shaping mask to also constrain the spatial signature of the interference, so that the tolerable interference power can increase without compromising the PU performance. We design two different algorithms for the computation of the spatial shaping matrices, and also addressed the optimization of different secondary networks under the foregoing constraints. Our results show that spatial shaping provides remarkable improvements to the secondary network in comparison to interference power constraints. We finally consider the TWRC with two single-antenna source nodes, whose link is too weak to establish direct communication, and multiple multi-antenna relays that follow the amplify-and-forward (AF) protocol. Differently from the two previous scenarios, the relays can access the user data, so that joint processing can be performed. However, the CSI knowledge is still a critical point. To avoid the global CSI requirement of the optimal relaying strategy, we contribute with a distributed algorithm that permits achieving rate region points that are very close to the optimal rate region boundary.