Cooperative ARQ With Relay Selection: An Analytical Framework Using Semi-Markov Processes

An important building block in cooperative diversity is relay selection, which has to ensure that a well-suited node is employed as a relay. The required coordination among nodes causes signaling overhead, which in turn can significantly devalue the performance benefits gained by cooperative diversity. A relay update policy defines when a new relay is selected; it can balance the tradeoff between performance and overhead. This tradeoff is studied using mathematical methods. We consider three relay selection schemes, i.e., permanent, reactive, and adaptive, which have different relay update rules. We develop an analytical framework using semi-Markov processes to evaluate the throughput and energy efficiency of cooperative automatic repeat request (ARQ) protocols in time-correlated multipath fading channels. Results reveal potential performance gains of different selection schemes under various conditions. The reactive and adaptive schemes make use of better suited relays due to frequent selections. If their selection overhead, however, is significant, a permanent relay can achieve higher throughput due to negligible overhead. The impact of temporal correlation of fading channels on throughput and energy efficiency is also shown. These insights can be applied for development of cooperative communication protocols.