With the evolution of integrated circuit (IC) techniques, full-duplex has been considered as a potential remedy to double the spectrum efficiency of relay networks as only one channel is needed per two hops, as compared with the conventional half-duplex relays. However, the inclusion of the full-duplex feature at relays inevitably suffers from a self-interference problem, thereby deteriorating the end-to-end system performance, and it solicits more studies to render full-duplex feasible. In addition, it is critical to provide convenient and perpetual energy to low-powered wireless relays via wirelessly charging, especially in the era of Internet of Things (IoT). A major concern for applying energy harvesting is that there exists a performance tradeoff between wireless energy and information transfer due to the limited resource at the relays. The goal of this project is to address these challenges by investigating full-duplex relay networks with or without rechargeable wireless relays, in which amplify-and-forward (AF) and selective decode-and-forward (SDF) transmission protocols are both taken into consideration. Instead of simply treating the self-interference as noise, we attempt to effectively utilize the self-interference in the network to facilitate both signal decoding and energy harvesting. For the full-duplex AF relay network, we theoretically analyze the capacity outage performance by fully exploiting the data information embedded in the self-interference at the destination node. We then theoretically revisit the performance of the full-duplex rechargeable AF relay network with the harvested energy served as relay transmit power, where a power splitter with an adjustable ratio is further applied at the relay to divide the received signal into two portions for signal decoding and energy harvesting purposes. For the full-duplex SDF relay network, we theoretically analyze the performance by capturing the impact of the self-interference on the symbol error rate. As an extension, we theoretically analyze the symbol error rate and capacity outage performance with the rechargeable relay. With the theoretical analyses, we jointly optimize the system parameters, e.g., transmit power, power splitter ratio, and relay deployment, by minimizing the capacity outage or symbol error rate. Finally, computer simulation will be conducted to verify the effectiveness of the proposed schemes and to show the system performance.