Article Preview
Top1. Introduction
Cognitive radios have been proposed to be the next generation wireless devices that can share underutilized spectrum (Haykin, 2005; Hossain & Bhagava, 2007; Mitola & Maguire, 1999). Spectrum sensing and dynamic spectrum access are the main principles of cognitive radios. In spectrum sensing, cognitive radio users (secondary users - SU) sense the spectrum of licenced users (primary users - PU) to detect and utilize spectrum holes within the PUs’ spectrum. The cognitive radio networks adopt a hierarchical access structure by considering the PUs as the legacy spectrum holders, as well as the SUs as the unlicensed users.
The challenge for a reliable sensing algorithm is to identify suitable transmission opportunities without compromising the integrity of the PUs (Almalfouh & Stuber, 2012; Pang & Scutari, 2013). The efficiency of the employed spectrum sensing technique plays a key role in maximizing the cognitive radio network throughput, while protecting the PUs from interference. The popular criterion in designing sensing techniques is to minimize the probability of false alarm as low as possible (Almalfouh & Stuber, 2012; Pang & Scutari, 2013). In addition, in order to limit the probability of interfering with PUs, it is desirable to keep the missed detection probability as low as possible. The sensing time is the tradeoff factor between the quality and the speed of sensing. Increasing the sensing times allows to have both low false alarm and low missed detection probabilities, but reduces the time available for transmissions which results in low throughputs of SUs. Another tradeoff factor between the false alarm and the missed detection probabilities includes the detection thresholds. Low detection thresholds result in high false alarm probability and low missed detection probability and vice versa. Thus, to maximize the throughput of SUs, a joint optimization of the sensing and transmission parameters of the SUs is required. The optimization parameters include: sensing times, decision threshold vectors, and power allocation vectors.
The joint spectrum sensing design and power control for a single-channel point-to-point cognitive radio network has been introduced in (Almalfouh & Stuber, 2012). In that work, the authors formulated the joint sensing-duration design and power control problem as a two-stage stochastic program with recourse. Their numerical results show that the method obtains a good throughput for the SUs. However, only the sensing time parameter in the spectrum sensing process is considered as an optimization variable, while the decision threshold is prefixed. Furthermore, the obtained throughput is not optimal because of using a two-stage stochastic programming in which the first stage is for finding the optimal sensing time, and the second stage is for finding the optimal power.
A game theory approach for the joint sensing and power allocation optimization problem of a multiple-channel multiple-SU cognitive radio network has been proposed by Pang and Scutari (Pang & Scutari, 2013). In that work, a novel class of Nash problems has been introduced, where each SU considered as a player competes against other SUs to maximize his/her own opportunistic throughput by choosing jointly the sensing duration, the detection thresholds, and the power allocation vector over multichannel link. Several constraints included interference constraints, probability constraints, and power budget constraints have been used to setup the game. The resulting players’ optimization problems are so non-convex that it is very challenging to solve it in the traditional game theory framework. To deal with the non-convexity of the game, the Quasi-Nash Equilibrium is proposed to obtain a reasonable performance. However, using too many constraints makes the joint optimization problem less dynamic to obtain a global optimality.