Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

SmartUT: Receive Beamforming for Spectral Coexistence of NGSO Satellite Systems

Almoatssimbillah Saifaldawla, Eva Lagunas, Flor Ortiz, Abuzar B. M. Adam, Symeon Chatzinotas

TL;DR

This work targets downlink co-frequency interference mitigation for NGSO satellite coexistence by addressing the limitations of traditional adaptive beamformers that rely on CSI and covariance estimates. It proposes MambaBF, a self-supervised DL-based receive beamformer that directly outputs a beamforming weight vector $oldsymbol{w}_{ ext{MBF}}$ from array snapshots $oldsymbol{Y}$, without requiring CSI and with UT-side deployment using Mamba state-space model layers to capture spatial-temporal features. The method is trained offline using only the available snapshots and a custom loss $\\mathcal{L}_{ASINR}$ that optimizes SINR, and is then applied online to maximize SINR while nulling interferers. Simulations show MambaBF outperforms MVDR/SMI, MRC, and in many cases approaches ZF performance, especially under imperfect CSI and limited snapshots, highlighting its practical potential for robust, low-complexity interference mitigation in NGSO spectral coexistence.

Abstract

In this paper, we investigate downlink co-frequency interference (CFI) mitigation in non-geostationary satellites orbits (NGSOs) co-existing systems. Traditional mitigation techniques, such as Zero-forcing (ZF), produce a null towards the direction of arrivals (DOAs) of the interfering signals, but they suffer from high computational complexity due to matrix inversions and required knowledge of the channel state information (CSI). Furthermore, adaptive beamformers, such as sample matrix inversion (SMI)-based minimum variance, provide poor performance when the available snapshots are limited. We propose a Mamba-based beamformer (MambaBF) that leverages an unsupervised deep learning (DL) approach and can be deployed on the user terminal (UT) antenna array, for assisting downlink beamforming and CFI mitigation using only a limited number of available array snapshots as input, and without CSI knowledge. Simulation results demonstrate that MambaBF consistently outperforms conventional beamforming techniques in mitigating interference and maximizing the signal-to-interference-plus-noise ratio (SINR), particularly under challenging conditions characterized by low SINR, limited snapshots, and imperfect CSI.

SmartUT: Receive Beamforming for Spectral Coexistence of NGSO Satellite Systems

TL;DR

This work targets downlink co-frequency interference mitigation for NGSO satellite coexistence by addressing the limitations of traditional adaptive beamformers that rely on CSI and covariance estimates. It proposes MambaBF, a self-supervised DL-based receive beamformer that directly outputs a beamforming weight vector from array snapshots , without requiring CSI and with UT-side deployment using Mamba state-space model layers to capture spatial-temporal features. The method is trained offline using only the available snapshots and a custom loss that optimizes SINR, and is then applied online to maximize SINR while nulling interferers. Simulations show MambaBF outperforms MVDR/SMI, MRC, and in many cases approaches ZF performance, especially under imperfect CSI and limited snapshots, highlighting its practical potential for robust, low-complexity interference mitigation in NGSO spectral coexistence.

Abstract

In this paper, we investigate downlink co-frequency interference (CFI) mitigation in non-geostationary satellites orbits (NGSOs) co-existing systems. Traditional mitigation techniques, such as Zero-forcing (ZF), produce a null towards the direction of arrivals (DOAs) of the interfering signals, but they suffer from high computational complexity due to matrix inversions and required knowledge of the channel state information (CSI). Furthermore, adaptive beamformers, such as sample matrix inversion (SMI)-based minimum variance, provide poor performance when the available snapshots are limited. We propose a Mamba-based beamformer (MambaBF) that leverages an unsupervised deep learning (DL) approach and can be deployed on the user terminal (UT) antenna array, for assisting downlink beamforming and CFI mitigation using only a limited number of available array snapshots as input, and without CSI knowledge. Simulation results demonstrate that MambaBF consistently outperforms conventional beamforming techniques in mitigating interference and maximizing the signal-to-interference-plus-noise ratio (SINR), particularly under challenging conditions characterized by low SINR, limited snapshots, and imperfect CSI.
Paper Structure (13 sections, 18 equations, 6 figures, 1 table)

This paper contains 13 sections, 18 equations, 6 figures, 1 table.

Table of Contents

  1. INTRODUCTION
  2. SYSTEM MODEL
  3. Channel and signal models
  4. Traditional beamforming techniques
  5. PROPOSED APPROACH
  6. Model Overview
  7. Core Mamba Layer (MmL).
  8. self-supervised Training Objective
  9. SIMULATION RESULTS
  10. Parameters and Data Generation
  11. Model Training
  12. Model Performance
  13. CONCLUSION AND FUTURE WORK

Figures (6)

  • Figure 1: Desired satellite downlink (green) with aggregated interference from a second independent NGSO constellation (red).
  • Figure 2: Systematic view of the proposed MambaBF approach, "n" refer to the training sample.
  • Figure 3: Mamba layer (MmL)
  • Figure 4: $\mathrm{ASINR}$ for multi-snapshot with (a) perfect CSI (left) and (b) imperfect CSI (right)
  • Figure 5: Beam nulling performance (representative case 1: perfect CSI)
  • ...and 1 more figures