Frontiers in Marine Science (Oct 2022)

Spatial-digital joint self-interference cancellation method for in-band full-duplex underwater acoustic communication

  • Yinheng Lu,
  • Yinheng Lu,
  • Yinheng Lu,
  • Xin Qing,
  • Xin Qing,
  • Xin Qing,
  • Chenlu Yang,
  • Chenlu Yang,
  • Chenlu Yang,
  • Yunjiang Zhao,
  • Songwen Wu,
  • Songwen Wu,
  • Songwen Wu,
  • Gang Qiao,
  • Gang Qiao,
  • Gang Qiao

DOI
https://doi.org/10.3389/fmars.2022.1015836
Journal volume & issue
Vol. 9

Abstract

Read online

The in-band full-duplex underwater acoustic communication (IBFD-UWAC) mode has twice the information throughput of the traditional half-duplex communication mode, significantly increasing the communication efficiency. Extracting the weak desired signal from the high-power self-interference signal without distortion remains a challenging problem in implementing IBFD-UWAC systems. This paper proposes a spatial-digital joint self-interference cancellation (SDSIC) method for IBFD-UWAC. We first perform spatial self-interference cancellation (SSIC) and propose an improved wideband constant-beamwidth beamformer to overcome the problem of direction- and array-dependent interference in IBFD-UWAC systems. Convex optimization is used to maintain a constant beam response in the main flap and cancel the self-interference signal from a fixed direction, thus increasing the signal-to-interference ratio of the desired signal. Subsequently, we perform digital self-interference cancellation (DSIC) on the residual self-interference signal, and propose a variable-step-size least-mean-squares algorithm based on the spatial noise threshold. This algorithm modifies the least-mean-squares step-size adjustment criterion according to the noise level after SSIC and the desired signal, resulting in better DSIC. A series of simulations are implemented in a hardware-in-the-loop platform to verify the practicality and real-time performance of the proposed SDSIC method. The results show that the self-interference signal power can be reduced by 41.5 dB using the proposed method, an improvement of 13.5 dB over the conventional SIC method.

Keywords