Selective cooperative decoding based on a hard-decision-aided log-likelihood ratio approximation

This study presents novel strategies to improve the decoding performance through base-station cooperation. In the proposed schemes, the soft output of the demapper at a serving base station (SBS) is replaced by a combination of log-likelihood ratio information, which is generated through demodulation by neighbouring base stations (NBSs) as well as the SBS. For the information used in the cooperation, soft information is first considered. Combining the soft information transmitted from all NBSs provides close-to-optimal performance, but it still requires a high amount of overhead to be passed over the backhaul links. In the proposed schemes, the SBS only cooperates with the reliable NBSs, which are identified on the basis of the magnitude of the metrics transmitted from the NBSs, to achieve a trade-off between performance and overhead. In addition, the use of hard information instead of soft information is also proposed to further reduce the backhaul overhead. In this scheme, the metric values as well as the hard information are exploited to generate the updated soft information for the channel decoder. In the simulation results, it is shown that the proposed schemes provide better error performance compared with the conventional schemes under limited backhaul capacities.