Trade-offs in analog sensing and communication in RF energy harvesting wireless sensor networks

Radio-frequency (RF) energy harvesting (EH) is an appealing solution for making wireless sensor networks (WSNs) self-reliant in terms of energy. We investigate the problem of sensing and estimation in a WSN for a practically motivated transmit and receive model. In it, noisy readings are communicated by multiple peak-power constrained RF EH sensor nodes in an analog manner using phase modulation to a fusion node, which uses the popular phase-locked loop (PLL) circuit for signal reception. For the time-sharing model, in which an EH sensor node alternately harvests energy and transmits data, and for a general class of stationary and ergodic RF EH processes, we present insightful expressions for the mean squared error (MSE) of the estimate at the fusion node, and optimal fraction of time a node harvests energy and optimal transmit power that minimize the MSE. Benchmarking with several digital schemes brings out the natural adaptability and efficacy of the considered scheme.