Signal Intelligence

Deep Learning for Signal Modulation Classification and Inference

English | 简体中文

---

Introduction

Signal Intelligence is an advanced framework for automatic modulation classification (AMC). Following a comprehensive refactor, the project now features a production-grade inference pipeline and modern project management tooling for high-integrity signal processing research.

---

Technical Specifications

Inference Pipeline

The project/inference.py module provides a clean interface for model deployment:

• Preprocessing: Automatic I/Q channel detection and shape normalization (1024 samples).
• Multi-Task Inference: Simultaneous prediction of modulation types, confidence levels, and symbol widths.
• Hardware Agnostic: Optimized for both CUDA and CPU-only environments.

Project Standards

project/
├── pyproject.toml      # Ruff & Metadata configuration
├── project/
│   ├── inference.py    # Production inference class
│   └── train.py        # Multi-task training orchestrator
└── tests/
    └── test_inference.py # Preprocessing & Pipeline validation

Installation & Execution

Prerequisites

• Python 3.8+
• PyTorch 1.9.0+

Execution

# Format and check code quality
ruff format . && ruff check .

# Run validation suite
pytest tests/

---

© 2026 AsaqeLee. Built for advanced signal processing research.