Power Allocation Scheme for Spectrally Precoded OFDMA Cognitive Radio

Spectral precoders in OFDM based cognitive radios can significantly reduce the sidelobe leakage in primary user's band. However, how to allocate power effectively in spectrally precoded OFDM systems is still an unexplored area. In this paper, we designed an orthogonal spectral precoder based on eigenvalue decomposition for minimizing the average interference leakage due to cognitive users in primary user's band and presented optimal power allocation strategy to maximize the capacity of the cognitive users while limiting the interference in primary user's band below a given threshold. For same power budget, the proposed optimal power allocation strategy achieves better capacity compared to equal power allocation. Moreover, in comparison to uncoded OFDM system (without spectral precoding), the proposed approach efficiently utilizes the available resources thereby improving the capacity of cognitive users significantly. The proposed approach is further extended for a more dynamic scenario where primary users occupying different sub-bands have distinct interference thresholds. For such cases, weighting factors are introduced whose values are chosen inversely proportional to the interference threshold in corresponding primary user's sub-band, we call it as proportional weighting approach. Simulation results show that, for distinct interference thresholds, the proportional weighting achieves better throughput compared to equal weighting.