Cognitive wireless personal area network for monitoring and control

In this paper, we propose and evaluate a cognitive personal area network (CPAN) suitable for wireless monitoring and control outside the Industrial, Scientific and Medical (ISM) band. Network is based on frequency hopping but the channels are selected on the basis of spectrum sensing performed by the nodes. Nodes balance their data collecting activity and spectrum sensing such that the duration of spectrum sensing done by a node is proportional to the number of packets previously transmitted by that node. We show that the operation of a node in this setup can be modeled as a gated exhaustive round-robin system with vacations, and analyze the performance of the system. The impact of the value of the penalty factor, which determines the duration of sensing activity with respect to transmission, on packet access delay and stability limits of the system is discussed in detail. Copyright © 2009 John Wiley & Sons, Ltd.