Markov modeling of energy harvesting Body Sensor Networks

The emerging paradigm of wearable and implantable medical sensors has enabled continuous and unobtrusive monitoring for patients and human subjects, allowing them to continue their normal activities, and yet be assured of immediate response in case of a detected health emergency. Energy harvesting has been proposed as a viable scheme for powering such sensors as periodic retrievals for battery replacements may not be feasible. The current state of the art in energy harvesting allows tapping into several physical and naturally existing sources, such as solar, wind, vibration, RF scavenging, among others. However, there is a lack of theoretical models that can predict future consumption and residual availability of energy in a sensor node equipped with multiple boards that can simultaneously operate on different types of sources. In this paper, we propose MAKERS, a Markov model based method to capture the energy states of such sensors. MAKERS allows detailed prediction of the probability of a node failing to detect an event owing to lack of energy, which is a key design consideration for body sensor sensors.