Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Neural Network-based Two-Dimensional Filtering for OTFS Symbol Detection

Jiarui Xu, Karim Said, Lizhong Zheng, Lingjia Liu

TL;DR

This work addresses OTFS symbol detection in high-mobility channels with limited OTA pilot data. It introduces a novel two-dimensional reservoir computing (2D-RC) architecture that performs online subframe detection directly in the delay-Doppler (DD) domain using a single neural network, leveraging a 2D windowing and 2D circular padding scheme to equalize the 2D circular channel. The training optimizes only the output weights and forget lengths, enabling efficient LS-based learning and reduced data requirements, while exploiting the OTFS structural knowledge to improve memory and performance. Empirical results show that 2D-RC yields substantial BER gains over the previous 1D-RC and competitive model-based detectors across modulation orders, without relying on perfect channel state information, which enhances practicality for real-world high-mobility deployments.

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 the limited over-the-air (OTA) pilot symbols are utilized for training. However, the previous RC-based approach does not design the RC architecture based on the properties of the OTFS system to fully unlock the potential of RC. This paper introduces a novel two-dimensional RC (2D-RC) approach for online symbol detection on a subframe basis in the OTFS system. The 2D-RC is designed to have a two-dimensional (2D) filtering structure to equalize the 2D circular channel effect in the delay-Doppler (DD) domain of the OTFS system. With the introduced architecture, the 2D-RC can operate in the DD domain with only a single neural network, unlike our previous work which requires multiple RCs to track channel variations in the time domain. Experimental results demonstrate the advantages of the 2D-RC approach over the previous RC-based approach and the compared model-based methods across different modulation orders.

Neural Network-based Two-Dimensional Filtering for OTFS Symbol Detection

TL;DR

This work addresses OTFS symbol detection in high-mobility channels with limited OTA pilot data. It introduces a novel two-dimensional reservoir computing (2D-RC) architecture that performs online subframe detection directly in the delay-Doppler (DD) domain using a single neural network, leveraging a 2D windowing and 2D circular padding scheme to equalize the 2D circular channel. The training optimizes only the output weights and forget lengths, enabling efficient LS-based learning and reduced data requirements, while exploiting the OTFS structural knowledge to improve memory and performance. Empirical results show that 2D-RC yields substantial BER gains over the previous 1D-RC and competitive model-based detectors across modulation orders, without relying on perfect channel state information, which enhances practicality for real-world high-mobility deployments.

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 the limited over-the-air (OTA) pilot symbols are utilized for training. However, the previous RC-based approach does not design the RC architecture based on the properties of the OTFS system to fully unlock the potential of RC. This paper introduces a novel two-dimensional RC (2D-RC) approach for online symbol detection on a subframe basis in the OTFS system. The 2D-RC is designed to have a two-dimensional (2D) filtering structure to equalize the 2D circular channel effect in the delay-Doppler (DD) domain of the OTFS system. With the introduced architecture, the 2D-RC can operate in the DD domain with only a single neural network, unlike our previous work which requires multiple RCs to track channel variations in the time domain. Experimental results demonstrate the advantages of the 2D-RC approach over the previous RC-based approach and the compared model-based methods across different modulation orders.

Paper Structure

This paper contains 15 sections, 17 equations, 6 figures.

Table of Contents

  1. Introduction
  2. System Model
  3. OTFS Transmitter and Receiver
  4. Channel
  5. Input-Output Relationship
  6. Problem formulation
  7. Introduced Approach -- 2D-RC
  8. Pre-processing
  9. 2D Windowing
  10. 2D Circular Padding
  11. Structure of 2D-RC
  12. Learning Algorithm
  13. Testing with 2D-RC
  14. Numerical Experiments
  15. Conclusion

Figures (6)

  • Figure 1: OTFS system diagram.
  • Figure 2: Pilot patterns. The green grids are filled with known pilot symbols. The green grid with a square marker denotes the spike pilot. The cross markers are guard symbols. The blank region represents data symbol positions.
  • Figure 3: 2D-RC Structure. For simplicity, the nonlinear function and the extended state are ignored here.
  • Figure 4: The 2D windowing process in 2D-RC.
  • Figure 5: BER comparison under QPSK.
  • ...and 1 more figures