Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Spatio-Temporal-Frequency Transformer Framework for Near-Field Target Recognition

Zongyao Zhao, Zhaolin Wang, Lincong Han, Jing Jin, Kaibin Huang

Abstract

A target recognition framework relying on near-field integrated sensing and communication (ISAC) systems is proposed. By exploiting the distance-dependent spatial signatures provided by the near-field spherical wavefront, high-accuracy sensing is realized in a bandwidth-efficient manner. A spatio--temporal--frequency (STF) transformer framework is introduced for target recognition using electromagnetic features found in the wireless channel response. In particular, a lightweight spatial encoder is employed to extract features from the antenna array for each frame and subcarrier. These features are then fused by a time-frequency transformer head with positional embeddings to model temporal dynamics and cross-subcarrier correlations. Simulation results demonstrate that strong target recognition performance can be achieved even with limited bandwidth resources.

A Spatio-Temporal-Frequency Transformer Framework for Near-Field Target Recognition

Abstract

A target recognition framework relying on near-field integrated sensing and communication (ISAC) systems is proposed. By exploiting the distance-dependent spatial signatures provided by the near-field spherical wavefront, high-accuracy sensing is realized in a bandwidth-efficient manner. A spatio--temporal--frequency (STF) transformer framework is introduced for target recognition using electromagnetic features found in the wireless channel response. In particular, a lightweight spatial encoder is employed to extract features from the antenna array for each frame and subcarrier. These features are then fused by a time-frequency transformer head with positional embeddings to model temporal dynamics and cross-subcarrier correlations. Simulation results demonstrate that strong target recognition performance can be achieved even with limited bandwidth resources.
Paper Structure (13 sections, 26 equations, 6 figures, 2 tables)

This paper contains 13 sections, 26 equations, 6 figures, 2 tables.

Table of Contents

  1. Introduction
  2. System and Signal Model
  3. Spatio--Temporal--Frequency Transformer Framework
  4. Data Representation and Preprocessing
  5. Network Architecture
  6. Spatial Encoding
  7. Time--Frequency Transformer Encoder
  8. Training and Optimization
  9. Experiments and Discussion
  10. Dataset and Experimental Setup
  11. Experimental Results
  12. Discussion
  13. Conclusion

Figures (6)

  • Figure 1: Illustration of the ISAC system with an extended target in the near-field regime.
  • Figure 2: Schematic of the proposed STF Transformer framework.
  • Figure 3: Test accuracy of STF and baselines under different sensing budgets.
  • Figure 4: Test accuracy surfaces of STF and baselines
  • Figure 5: Test accuracy versus $K_{\mathrm{sel}}$ with a fixed $N_p=4$.
  • ...and 1 more figures