Wideband PHY/MAC bandwidth aggregation optimization for cognitive radios

Unlike conventional radios, cognitive radios are to use channel occupancy information measured at the PHY layer and conveyed to the MAC layer to select suitable frequency bands to communicate. Many existing PHY/MAC decision-making strategies, however, assume that the cognitive radio belongs to a particular radio network and its communication capabilities are limited to the protocols supported by that radio network. In this paper, we propose a cognitive radio PHY/MAC decision-making strategy that may simultaneously utilize multiple radio networks across a wide spectrum band. The whole spectrum range is assumed to be divided into several sub-bands in performing spectrum sensing. Each of the sub-bands may have an arbitrary bandwidth, depending on the spectrum sensing capability of the cognitive radio. In this paper, we derive an optimal wideband bandwidth aggregation (BAG) strategy for the energy and frequency efficient communication problem: a multi-objective optimization problem is formulated, one objective is the communication throughput of the mobile cognitive radio device and the other one is energy consumption of the device. The proposed multi-objective optimization problem takes into account the essential practical issues including imperfect spectrum sensing, time varying channel coefficients, hardware reconfiguration time delay, hardware reconfiguration power consumptions, and communication power consumptions. The optimal BAG strategy is solved using a combination of the Hungarian algorithm and convex optimization. In this paper, we show that by self-adjusting the weighting coefficients of two objectives, the cognitive radio may achieve autonomous operation. The formulation can also be easily extended to multi-objective problems that have more than two objectives.