In a cognitive relay network, a secondary cooperative transmission, sharing the same spectrum band with N primary transmissions, is considered. Assuming imperfect channel state knowledge of the interference channels (from secondary transmitters to primary receivers) at the secondary nodes, we investigate the problem of secondary achievable rate maximization under a constraint on the interference caused at the primary receivers. Different scenarios are defined based on the accuracy of estimating the interference channels at the primary receivers and the number of feedback quantization bits.
For each scenario, we derive expressions for both the secondary achievable rate and the primary interference probability. Then, we investigate the associated power allocation problem and provide a sub-optimal solution. Finally, analytical and simulation results illustrate the efficiency of the provided solution in terms of secondary achievable rate and primary interference probability.