# Radar HRRP Sequence Target Recognition Based on a Lightweight Spatiotemporal Fusion Network

## TL;DR

This paper introduces a lightweight radar target recognition method that improves accuracy and efficiency for real-time applications.

## Contribution

A novel lightweight spatiotemporal fusion network for HRRP sequence recognition with improved robustness and efficiency.

## Key findings

- The proposed method outperforms existing approaches on MSTAR and CVDomes datasets.
- The lightweight design reduces computation and parameters for edge-device deployment.
- Adaptive focal loss with label smoothing improves performance on imbalanced classes.

## Abstract

High-resolution range profile (HRRP) sequence recognition in radar automatic target recognition faces several practical challenges, including severe category imbalance, degradation of robustness under complex and variable operating conditions, and strict requirements for lightweight models suitable for real-time deployment on resource-limited platforms. To address these problems, this paper proposes a lightweight spatiotemporal fusion-based (LSTF) HRRP sequence target recognition method. First, a lightweight Transformer encoder based on group linear transformations (TGLT) is designed to effectively model temporal dynamics while significantly reducing parameter size and computation, making it suitable for edge-device applications. Second, a transform-domain spatial feature extraction network is introduced, combining the fractional Fourier transform with an enhanced squeeze-and-excitation fully convolutional network (FSCN). This design fully exploits multi-domain spatial information and enhances class separability by leveraging discriminative scattering-energy distributions at specific fractional orders. Finally, an adaptive focal loss with label smoothing (AFL-LS) is constructed to dynamically adjust class weights for improved performance on long-tail classes, while label smoothing alleviates overfitting and enhances generalization. Experiments on the MSTAR and CVDomes datasets demonstrate that the proposed method consistently outperforms existing baseline approaches across three representative scenarios.