On the Capacity of Cognitive Radio Sensor Networks: From Physical Layer to Network Layer

Cognitive Radio has stirred great interest recently, with its tremendous potential to further exploit the scarce spectrum resource. Among the diverse research topics, the fundamental capacity issue deserves a thorough study since it serves as a principle for future implementation and performance evaluation. Existing researches on cognitive capacity mainly fall into two categories: channel capacity and network capacity. The former can accurately portray the pairwise cognitive channel capacity, but less address the case with multiple secondary users. The latter investigates the throughput of large-scale network with respect to its size and topology, while ignoring crucial factors in physical layer Recognizing their strengths and limitations, this paper aims to present a more comprehensive view of the cognitive capacity issue. First, we extend the information theoretic model to a special case in multi-user scenario, which well characterizes the capacity of cognitive multiple access channel (MAC). Second, we analyze network scaling by incorporating sensing errors, and show that physical layer features may affect the throughput performance. Last but not least, our analysis not only extends the existing researches but also sheds some light on the cross-layer study of cognitive capacity.