Stochastic modeling and analysis of wireless sensor nodes with hybrid storage systems

The notion of perpetual operation using energy harvesting sensor networks are usually crippled by the limited cycle life of the rechargeable batteries (RB). Though the idea of complementing the operation of RBs by using supercapacitors (SCs) is not completely new, very little has been done to effectively model the energy dynamics in such hybrid energy storage systems (HESS) of the future. The contribution of this paper is to capture these dynamics, thereby enabling an accurate prediction of the energy intake and consumption to design future energy harvesting systems and to quantify the performance of practical energy storage systems. Specifically, we propose a practical, dynamic energy model for a wireless node that is based on a four-dimensional Markov Chain. The proposed dynamic model emulates a quasi-birth-death (QBD) process, in which random and unpredictable sources of energy harvesting and routing requests (i.e., packet arrivals) are considered. The outcome is a compact analytical tool that is a first step in gaining insights into the long-term performance metrics and energy usage due to route participation and characteristics of an energy-constrained wireless node.