Self-interference Mitigation in Full-duplex Relays:Blind Adaptive Techniques and Optimal Linear Filter Design

Wireless communication systems need to satisfy the increasing demand for high data rates, low latency and ubiquitous connectivity. Besides, efficient resource management and a high spectral efficiency are also paramount. Multi-hop communication via relays is an attractive and cost-effective solution for cooperative communication and coverage extension. Full-duplex technology, i.e., simultaneous transmission and reception in the same frequency band, offers twice the spectral efficiency than its half-duplex counterpart, where transmission and reception take place in different time slots. Full-duplex technology makes possible for devices to operate directly on the waveform samples without synchronizing to the network, thus allowing to reduce the end-to-end latency and posing itself as an ideal solution for wireless systems. However, the implementation of a full-duplex device has several practical issues. Of utmost importance is the self-interference, which causes the transmitted signal to interfere with the received signal, leading to a severe drop of the signal-to-noise ratio. To fully take advantage of the full-duplex technology, self-interference must be mitigated. This dissertation addresses the issue of self-interference mitigation for different relay protocols and configurations. The first theme deals with adaptive blind algorithms for self-interference cancellation. Assuming no available training information or synchronization with the network, we propose algorithms that yield residual self-interference below noise level. The second theme considers the optimization of a full-duplex relay link subjected to limited dynamic range, i.e., impairments at transmission and reception sides. Considering available channel state information, we design optimal linear filters at each node under the maximum signal-to-noise ratio and the minimum mean square error criteria.