[Paper Review] A Comprehensive Survey of Machine Learning Applied to Radar Signal Processing

Ping Lang, Xiongjun Fu+5 more|arXiv (Cornell University)|Sep 29, 2020

Advanced SAR Imaging Techniques565 references34 citations

TL;DR

This survey provides a structured review of traditional ML and deep learning methods applied to radar signal processing, covering emitter recognition, SAR/ISAR image processing, anti-jamming, and related challenges and directions.

ABSTRACT

Modern radar systems have high requirements in terms of accuracy, robustness and real-time capability when operating on increasingly complex electromagnetic environments. Traditional radar signal processing (RSP) methods have shown some limitations when meeting such requirements, particularly in matters of target classification. With the rapid development of machine learning (ML), especially deep learning, radar researchers have started integrating these new methods when solving RSP-related problems. This paper aims at helping researchers and practitioners to better understand the application of ML techniques to RSP-related problems by providing a comprehensive, structured and reasoned literature overview of ML-based RSP techniques. This work is amply introduced by providing general elements of ML-based RSP and by stating the motivations behind them. The main applications of ML-based RSP are then analysed and structured based on the application field. This paper then concludes with a series of open questions and proposed research directions, in order to indicate current gaps and potential future solutions and trends.

Motivation & Objective

• Provide a systematic overview of ML-based radar signal processing (RSP) techniques across traditional ML and deep learning.
• Analyze applications in radar emitter recognition, SAR/ISAR image processing, anti-jamming, and waveform design.
• Identify gaps, challenges, and future research directions to guide researchers and practitioners.

Proposed method

• Survey and synthesis of over 600 papers published around 2015–2020 from IEEE Xplore, Web of Science, and dblp.
• Structured analysis of ML-based RSP techniques by application field and ML paradigm (traditional ML and DL).
• Discussion of motivations, concepts, and implications of enabling intelligent algorithms in RSP.
• Comparison of ML models across RSP tasks and identification of open questions and trends.

Experimental results

Research questions

• RQ1 What ML and DL techniques have been applied to radar signal processing and in what problem settings?
• RQ2 What are the main application areas and how do ML approaches perform across them?
• RQ3 What are the current gaps, limitations, and open questions guiding future ML-based RSP research?
• RQ4 What future trends and research directions are most promising for ML in radar sensing?

Key findings

• ML-based RSP spans traditional models (SVMs, decision trees, random forests, boosting, XGBoost) and deep learning models (CNNs, RNNs, DBNs, AEs, GANs).
• Applications include radar emitter recognition, SAR/ISAR image processing, anti-jamming, waveform design, and cognitive electronic warfare.
• DL methods offer powerful feature learning for radar imagery and signals, while traditional ML provides robust classifiers for emission recognition and related tasks.
• The survey highlights open questions and future directions such as data scarcity, model interpretability, and integration of multi-representation learning approaches.