Performance Analysis of Optimum Combining under Multiple Primary Interferers in Underlay Cognitive Radio

Abstract

Cognitive radio (CR) allows secondary users to share the same spectrum with primary users. The spectrum sharing should be carried out in such a way that the interference generated by the secondary user is held below a minimum threshold referred as interference temperature. The capacity of the Secondary user degrades due to this constraint .Therefore, to improve the performance of secondary user, diversity techniques can be used at the receiver of secondary user. Moreover, the performance of the secondary user is affected by the interference from the transmitter of primary user. Therefore, the diversity techniques should also be able to reduce the impact of interference in addition to reduce the fading. In this dissertation work, the performance of underlay CR system is studied with optimum combining (OC) at the SU receiver under the impact of interference from multiple PU-Txs in flat Rayleigh fading channels. The number of primary user’s transmitters 𝐿𝑡 and antennas at SU receiver 𝐾𝑟 are related as 𝐿𝑡 ≥ 𝐾𝑟. An expression is derived for the approximated probability density function of maximum SIR at secondary user receiver output. Using this density function, closed form expressions are obtained for the average post processing SIR, ergodic capacity ,outage probability and average bit error rate of the CR-OC system taking into consideration peak interference power constraint denoted as 𝑄𝑝 at primary user receiver. Performance of proposed system is compared with the maximal ratio combining in cognitive radio (CR-MRC). Analytical results for CR-OC system are validated through Monte Carlo simulations. Based on achieved results, it is shown that OC is significantly better than MRC even when the number of primary interferers exceeds the number of antennas at secondary user receiver. Also CR-OC with two more interferers gives the same performance as CR-MRC. In order to achieve average bit error rate of 10−1 for 𝐿𝑡=3 for both the systems, CR-OC requires 5dB less power as compared to CR-MRC. In addition, with 𝐿𝑡=6, for an average bit error rate of 10−0.9, OC leads to a power saving of nearly 1.5dB over MRC