Primary user activity modeling using multi-term parameter estimation in cognitive radio systems

Primary user (PU) activity modeling is a technique used to produce more information at the secondary user (SU) during its spectrum sensing process. With the PU activity information, the SU is able to obtain a better decision in selecting a vacant channel for its communication. However, the long-term parameter estimation based activity modeling is seen to be missing the valuable spectrum opportunities under burst traffic conditions while the short-term parameter estimation based PU activity modeling techniques can capture these opportunities with an added complexity. At the same time, poor correlation among traffic at different time intervals can dramatically reduce the accuracy in short-term modeling. On the other hand, the actual PU behavior can be shadowed from the SU due to the practical channel impairments. These give rise to a situation that neither the short-term nor the long-term parameter estimation based activity modeling are accurate under all conditions. This paper proposes a multi-term parameter estimation based PU activity modeling scheme which produces a higher spectrum sensing efficiency while retaining the simplicity. Furthermore, the proposed scheme produces an efficient way of primary user activity modeling under realistic channel conditions. The simulation results verify the superior modeling accuracy versus the complexity trade-off of the proposed scheme when compared to the existing modeling techniques.