Improved Energy-Based Signal Detection for Ambient Backscatter Communications

TL;DR

This paper introduces two novel detectors, JCED and IED, that significantly improve signal detection performance in ambient backscatter communications over traditional energy detectors, especially under uncertain channel conditions.

Contribution

The paper proposes two new detection methods, JCED and IED, that outperform traditional energy detection in AmBC systems under generalized noise models and various antenna configurations.

Findings

Both detectors outperform traditional energy detectors in detection probability.

JCED and IED show significant BER improvements with multiple antennas.

Performance gains are quantified under different false alarm rates and interference conditions.

Abstract

In ambient backscatter communication (AmBC) systems, passive tags connect to a reader by reflecting an ambient radio frequency (RF) signal. However, the reader may not know the channel states and RF source parameters. The traditional energy detector (TED) appears to be an ideal solution. However, it performs poorly under these conditions. To address this, we propose two new detectors: (1) A joint correlation-energy detector (JCED) based on the first-order correlation of the received samples and (2) An improved energy detector (IED) based on the p-th norm of the received signal vector. We compare the performance of IED and TED under the generalized noise model using McLeish distribution and derive a general analytical formula for the area under the receiver operating characteristic (ROC) curves. Based on our results, both detectors outperform TED. For example, the probability of detection with a false alarm rate of 1\% for JCED and IED is 22.97\% and 5.41\% higher, respectively, compared to TED for a single-antenna reader. Using the direct interference cancellation (DIC) technique, these gains are 34.92\% and 3.7\%, respectively. With a four-antenna array at the reader and a 5\% false alarm rate, the JCED shows a significant BER improvement of 28.68\% without DIC and 48.21\% with DIC.