Design of Ambient Backscatter Training for Wireless Power Transfer

TL;DR

This paper introduces a novel ambient backscatter training method for wireless power transfer that enhances energy harvesting efficiency in low-power IoT devices by eliminating interference without explicit channel estimation.

Contribution

It proposes a new training sequence design for ambient backscatter WPT that cancels interference and models the system analytically considering realistic fading and energy harvesting models.

Findings

Full interference cancellation when ambient symbol duration is known

Robustness to timing offset mismatches

Harvesting of tens to hundreds of microwatts in low-power IoT scenarios

Abstract

Wireless power transfer (WPT) using energy beamforming is a promising solution for low power Internet of Things (IoT) devices. In this work, we consider WPT from an energy transmitter (ET) employing retrodirective WPT using a large phased antenna array to an energy receiver (ER) capable of ambient backscatter. The advantage of retrodirective WPT is that no explicit channel estimation is needed at the ET and the use of ambient backscattering eliminates the need for active transmission at the ER. We propose a training sequence design, i.e., pattern of varying the reflection coefficient at the ER, to eliminate the direct-link interference from the ambient source. We show that when the ambient symbol duration is known, the ambient interference is fully cancelled by the proposed design. We analytically model the system and find the average harvested power at the ER considering Nakagami-$m$ fading channels and non-linear energy harvesting model. Our results clearly show that the proposed solution is robust to a small timing offset mismatch at the correlator. When interference from undesired neighbouring sources in the ambient environment is not significant, the ER can successfully harvest tens to hundreds of $\mu$W of power, which is an important improvement for low-power IoT devices.