The fusion of Wireless Sensor Networks (WSNs) and Cognitive Radio Networks (CRNs) into Cognitive Radio Sensor Networks (CRSNs) is quite an attractive proposal, because it allows a distributed set of low-powered sensor nodes to opportunistically access spectrum bands that are underutilized by their licensed owners (called primary users (PUs)). In addition, when the PUs are actively transmitting in their own bands, sensor nodes can switch to energy harvesting mode to obtain their energy needs (for free), to achieve almost perpetual life. In this work, we present a novel and fully distributed MAC protocol, calledS-LEARN, that allows sensor nodes in a CRSN to entwine their RF energy harvesting and data transmission activities, while intelligently addressing the issue of disproportionate difference between the high power necessary for the node to transmit data packets and the small amount of power it can harvest wirelessly from the environment. The presented MAC protocol can improve both the network throughput and total harvested energy, while being robust to changes in the network configuration. Moreover, S-LEARN can keep the cost of the system low, and it avoids the pitfalls from which centralized systems suffer.

IEEE 802.22 wireless regional area network (WRAN) is a cognitive radio-based network. WRANs are intended to be deployed by different service providers and designed to opportunistically utilize the unused TV bands. WRANs have to self-coexist with other overlapped WRANs in a distributed manner. Therefore, every service provider tries to acquire a band free of interference from others to satisfy a required quality of service. This self-coexistence problem is one of the major challenges in WRAN. In this paper, we propose a Markov-based distributed approach for mitigating this problem. We model the problem as an absorbing discrete-time Markov chain. In this model, if two or more overlapped WRANs select the same band, then each one should either stay or switch to another band according to a certain switching probability. This process continues until each one of the WRANs finds an interference-free band. In this case, the Markov chain reaches the absorbing state. This model is employed to find the optimal switching probability, which in turn minimizes the time required to reach the absorbing state. The switching probability is numerically found as a function of the number of overlapped WRANs and available bands. Extensive simulation has been conducted to validate our numerical results.