Many wireless systems have leveraged the cooperation of nodes to improve network capacity. However, these approaches have left uncertainty over the performance gain of spatial diversity either with protocol overheads or unrealistic assumptions. Also more practical relay selection algorithm is required to be simple and totally distributed. In this paper, a CSMA-liked queue length based algorithm is proposed that randomly schedules ARQ relays, and involves other sessions to create network coding opportunities in erroneous hyperedges. Through partial data recovery, we exploit the probabilistic nature of wireless reception for joint cooperation, scheduling and ARQ to provide relevant gains (e.g., transmission reliability and opportunity to access to the channel) for both the source and cooperating nodes.

Keywords