Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

2D-RC: Two-Dimensional Neural Network Approach for OTFS Symbol Detection

Jiarui Xu, Karim Said, Lizhong Zheng, Lingjia Liu

TL;DR

This work advances OTFS symbol detection by embedding domain knowledge of the 2D delay-Doppler interaction into a reservoir-computing framework, yielding the 2D-RC detector. By employing 2D circular padding and a 2D filtering structure, the method operates in the DD domain with a single neural network, enabling online subframe-based learning without explicit CSI. Empirical results show substantial gains over previous RC approaches and model-based detectors across RCP-OTFS and CP-OTFS, for QPSK and 16QAM, and under LDPC coding, while maintaining competitive complexity. The approach offers improved generalization and adaptability to OTFS variants, making it attractive for practical high-mobility wireless systems.

Abstract

Orthogonal time frequency space (OTFS) is a promising modulation scheme for wireless communication in high-mobility scenarios. Recently, a reservoir computing (RC) based approach has been introduced for online subframe-based symbol detection in the OTFS system, where only a limited number of over-the-air (OTA) pilot symbols are utilized for training. However, this approach does not leverage the domain knowledge specific to the OTFS system to fully unlock the potential of RC. This paper introduces a novel two-dimensional RC (2D-RC) method that incorporates the domain knowledge of the OTFS system into the design for symbol detection in an online subframe-based manner. Specifically, as the channel interaction in the delay-Doppler (DD) domain is a two-dimensional (2D) circular operation, the 2D-RC is designed to have the 2D circular padding procedure and the 2D filtering structure to embed this knowledge. With the introduced architecture, 2D-RC can operate in the DD domain with only a single neural network, instead of necessitating multiple RCs to track channel variations in the time domain as in previous work. Numerical experiments demonstrate the advantages of the 2D-RC approach over the previous RC-based approach and compared model-based methods across different OTFS system variants and modulation orders.

2D-RC: Two-Dimensional Neural Network Approach for OTFS Symbol Detection

TL;DR

This work advances OTFS symbol detection by embedding domain knowledge of the 2D delay-Doppler interaction into a reservoir-computing framework, yielding the 2D-RC detector. By employing 2D circular padding and a 2D filtering structure, the method operates in the DD domain with a single neural network, enabling online subframe-based learning without explicit CSI. Empirical results show substantial gains over previous RC approaches and model-based detectors across RCP-OTFS and CP-OTFS, for QPSK and 16QAM, and under LDPC coding, while maintaining competitive complexity. The approach offers improved generalization and adaptability to OTFS variants, making it attractive for practical high-mobility wireless systems.

Abstract

Orthogonal time frequency space (OTFS) is a promising modulation scheme for wireless communication in high-mobility scenarios. Recently, a reservoir computing (RC) based approach has been introduced for online subframe-based symbol detection in the OTFS system, where only a limited number of over-the-air (OTA) pilot symbols are utilized for training. However, this approach does not leverage the domain knowledge specific to the OTFS system to fully unlock the potential of RC. This paper introduces a novel two-dimensional RC (2D-RC) method that incorporates the domain knowledge of the OTFS system into the design for symbol detection in an online subframe-based manner. Specifically, as the channel interaction in the delay-Doppler (DD) domain is a two-dimensional (2D) circular operation, the 2D-RC is designed to have the 2D circular padding procedure and the 2D filtering structure to embed this knowledge. With the introduced architecture, 2D-RC can operate in the DD domain with only a single neural network, instead of necessitating multiple RCs to track channel variations in the time domain as in previous work. Numerical experiments demonstrate the advantages of the 2D-RC approach over the previous RC-based approach and compared model-based methods across different OTFS system variants and modulation orders.
Paper Structure (28 sections, 44 equations, 17 figures, 2 tables)

This paper contains 28 sections, 44 equations, 17 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Preliminaries -- Reservoir computing
  3. Pre-processing
  4. Windowing
  5. Padding
  6. Structure of 1D-RC
  7. Learning algorithm
  8. Testing with 1D-RC
  9. System Model
  10. OTFS transmitter and receiver
  11. Channel
  12. Variants of OTFS system
  13. RCP-OTFS system
  14. CP-OTFS system
  15. Problem formulation
  16. ...and 13 more sections

Figures (17)

  • Figure 1: 1D-RC Structure. For simplicity, the extended state and nonlinear function are ignored here. In the figure, the target output is a sequence with $N_o=1$.
  • Figure 2: The windowing process in 1D-RC.
  • Figure 3: OTFS system diagram.
  • Figure 4: OTFS system variants.
  • Figure 5: Pilot patterns. (a) Blockwise pilot pattern. (b) Spike pilot pattern. The green grids are filled with known pilot symbols. The green grid with a square marker denotes the spike pilot. The cross marker represents the guard symbols. The blank region represents the data symbol position.
  • ...and 12 more figures