Square Deviation Based Symbol-Level Selection for Virtual Full-Duplex Relaying Networks.
Jiancao Hou,Sandeep Narayanan,Na Yi,Yi Ma,Mohammad Shikh-Bahaei
DOI: https://doi.org/10.1109/nof.2017.8251244
2017-01-01
Abstract:In this paper, a symbol-level selective transmission for virtual full-duplex (FD) relaying networks is proposed, which aims to mitigate error propagation effects and improve system spectral efficiency. The idea is to allow two half-duplex relays, mimicked as FD relaying, to alternatively serve as transmitter and receiver to forward the source's messages. In this case, each relay predicts the correctly decoded symbols of its received frame, based on the generalized square deviation method, and discard the erroneously decoded symbols, resulting in fewer errors being forwarded to the destination. Then, a modified maximum a posteriori receiver at the destination is provided to eliminate the inter-frame interference and identify the positions of discarded symbols from the relays. In addition, the diversity-multiplexing trade-off (DMT) for our proposed scheme is also analysed. It is found that our proposed scheme outperforms the conventional selective decode-and-forward (S-DF) relaying schemes, such as cyclic redundancy check based S-DF and threshold based S-DF, in terms of DMT. Moreover, the bit-error-rate performance are also simulated to confirm the DMT results.