Full-duplex communication is a promising technique which guarantees an enhanced spectral efficiency in modern 5G wireless communications. In this technique, same set of frequency channels is used for simultaneous uplink and downlink signal transmissions and hence is termed as full-duplex (FD) communications or full-duplex radios. However, a major shortcoming of this technique is the presence of self-interference (SI), which arises due to the presence of both transmitters and receivers in close proximity and in fact several solutions have been proposed to mitigate it. In this paper, we give a new insight on the applicability of neural networks in solving (linear and nonlinear) SI problems using hybrid cancelations.
Keywords
- Full-duplex radios
- Self-interference
- Hybrid cancelation
- Neural networks