Performance Characterization of Energy Harvesting Sensors with Limited Buffer and Battery Capacities

Energy harvesting is a key enabling technology to prolong the lifetime of wireless sensor devices, while a sink node should be properly placed to efficiently gather sensor data. In our model, sensors distributed randomly based on a Matern cluster process harvest energy from a Poisson point process of base stations and use this energy for their transmissions to a stationary aerial sink such as a tethered drone. Moreover, each sensor is assumed to have a battery with multiple energy levels and a packet buffer with a finite capacity. We aim to comprehensively characterize the harvesting and outage performance of radio frequency energy harvesting sensor devices with an aerial receiver. We derive the ambient power at a sensor device using a moment generating function based approach, and the harvested energy is characterized. For the scheme where a sensor transmits whenever the buffer becomes full irrespective of the battery level, we derive the interrelated steady-state probabilities of the different battery states and packet buffer levels using Markov chains. Finally, the outage probability is derived when a sensor transmits its packets to an aerial sink node. Our numerical results illustrate that the numbers of energy and buffer states significantly impact the energy harvesting performance and the overall outage.