Dynamic Power Optimization for Secondary Wearable Biosensors in
E-Healthcare Leveraging Cognitive WBSNs with Imperfect Spectrum Sensing
Abstract
Abstract The integration of cognitive radio with e-healthcare systems
assisted by wireless body sensor networks (WBSNs) has been regarded as
an enabling approach for a new generation of pervasive healthcare
services, to provide differentiated quality of service requirements and
avoid harmful electromagnetic interference to primary medical devices
(PMDs) over the crowded radio spectrum. Due to the sharing spectrum
bands with PMDs in e-healthcare scenario using cognitive WBSNs (CWBSNs),
efficient transmit power control and optimization strategies for
resource-constrained secondary wearable biosensors (SWBs) play a key
role in controlling the inter-network interference and improving the
energy efficiency. This paper investigates the problem of dynamic power
optimization for SWBs in e-healthcare leveraging CWBSNs with practical
limitations, e.g., imperfect spectrum sensing and quality of
physiological data sampling. We develop a distributed optimization
framework of dynamic power optimization via the theory of differential
game, by jointly considering utility maximization and quality of
physiological data sampling for every SWB, while satisfying the
evolution law of energy consumption in SWB’s battery. With the
non-cooperation and cooperation relations for all SWBs in mind, we
transform the differential game model into two subproblems, namely,
utility maximization problem and total utility maximization problem.
Utilizing Bellman’s dynamic programming, we derive a non-cooperative
optimal solution for power optimization as a Nash equilibrium point for
the utility maximization problem posed by competitive scenario. By
exploiting Pontryagin’s maximum principle, a cooperative optimal
solution is obtained for the total utility maximization problem, wherein
all SWBs fully cooperate to obtain the highest total utilities. Building
upon the analytical results, the actual utility distributed to each SWB
is compared between the non-cooperative and cooperative schemes.
Extensive simulations show that the proposed optimization framework is
indeed an efficient and practical solution for power control compared
with the benchmark algorithm.