Ultralight Signal Classification Model for Automatic Modulation Recognition

TL;DR

This paper introduces an ultralight hybrid neural network for automatic modulation recognition, optimized for edge devices, achieving high accuracy with minimal data and computational resources.

Contribution

The work presents a novel, resource-efficient neural network model tailored for real-time signal classification on edge devices, with robust performance in noisy environments.

Findings

Achieves 96.3% accuracy at 0 dB SNR

Operates effectively with fewer than 100 samples per class

Reduces computational overhead significantly

Abstract

The growing complexity of radar signals demands responsive and accurate detection systems that can operate efficiently on resource-constrained edge devices. Existing models, while effective, often rely on substantial computational resources and large datasets, making them impractical for edge deployment. In this work, we propose an ultralight hybrid neural network optimized for edge applications, delivering robust performance across unfavorable signal-to-noise ratios (mean accuracy of 96.3% at 0 dB) using less than 100 samples per class, and significantly reducing computational overhead.