This book shows you how to adopt data-driven techniques for the problem of radar detection, both per se and in combination with model-based approaches. In particular, the focus is on space-time adaptive target detection against a background of interference consisting of clutter, possible jammers, and noise. It is a handy, concise reference for many classic (model-based) adaptive radar detection schemes as well as the most popular machine learning techniques (including deep neural networks) and helps you identify suitable data-driven approaches for radar detection and the main related issues. You’ll learn how data-driven tools relate to, and can be coupled or hybridized with, traditional adaptive detection statistics; understand fundamental concepts, schemes, and algorithms from statistical learning, classification, and neural networks domains. The book also walks you through how these concepts and schemes have been adapted for the problem of radar detection in the literature and provides you with a methodological guide for the design, illustrating different possible strategies. You’ll be equipped to develop a unified view, under which you can exploit the new possibilities of the data-driven approach even using simulated data. This book is an excellent resource for Radar professionals and industrial researchers, postgraduate students in electrical engineering and the academic community.
Author(s): Angelo Coluccia
Series: Artech House Radar Library
Publisher: Artech House
Year: 2022
Language: English
Pages: 234
City: Boston
Adaptive Radar Detection Model-Based, Data-Driven, and Hybrid Approaches
Contents
Preface
Acknowledgments
1
Model-Based Adaptive Radar Detection
1.1 Introduction to Radar Processing
1.1.1 Generalities and Basic Terminology of Coherent Radars
1.1.2 Array Processing and Space-Time Adaptive Processing
1.1.3 Target Detection and Performance Metrics
1.2 Unstructured Signal in White Noise
1.2.1 Old but Gold: Basic Signal Detection and the Energy Detector
1.2.2 The Neyman–Pearson Approach
1.2.3 Adaptive CFAR Detection
1.2.4 Correlated Signal Model in White Noise
1.3 Structured Signal in White Noise
1.3.1 Detection of a Structured Signal in White Noise and Matched Filter
1.3.2 Generalized Likelihood Ratio Test
1.3.3 Detection of an Unknown Rank-One Signal in White Noise
1.3.4 Steering Vector Known up to a Parameter and Doppler Processing
1.4 Adaptive Detection in Colored Noise
1.4.1 One-Step, Two-Step, and Decoupled Processing
1.4.2 General Hypothesis Testing Problem via GLRT: A Comparison
1.4.3 Behavior under Mismatched Conditions: Robustness vs Selectivity
1.4.4 Model-Based Design of Adaptive Detectors
1.5 Summary
References
2 Classification Problems and Data-Driven Tools
2.1 General Decision Problems and Classification
2.1.1 M-ary Decision Problems
2.1.2 Classifiers and Decision Regions
2.1.3 Binary Classification vs Radar Detection
2.1.4 Signal Representation and Universal Approximation
2.2 Learning Approaches and Classification Algorithms
2.2.1 Statistical Learning
2.2.2 Bias-Variance Trade-Off
2.3 Data-Driven Classifiers
2.3.1 k-Nearest Neighbors
2.3.2 Linear Methods for Dimensionality Reduction and Classification
2.3.3 Support Vector Machine and Kernel Methods
2.3.4 Decision Trees and Random Forests
2.3.5 Other Machine Learning Tools
2.4 Neural Networks and Deep Learning
2.4.1 Multilayer Perceptron
2.4.2 Feature Engineering vs Feature Learning
2.4.3 Deep Learning
2.5 Summary
References
3
Radar Applications of Machine Learning
3.1 Data-Driven Radar Applications
3.2 Classification of Communication and Radar Signals
3.2.1 Automatic Modulation Recognition and Physical-Layer Applications
3.2.2 Datasets and Experimentation
3.2.3 Classification of Radar Signals and Radiation Sources
3.3 Detection Based on Supervised Machine Learning
3.3.1 SVM-Based Detection with Controlled PFA
3.3.2 Decision Tree-Based Detection with Controlled PFA
3.3.3 Revisiting the Neyman–Pearson Approach
3.3.4 SVM and NN for CFAR Processing
3.3.5 Feature Spaces with (Generalized) CFAR Property
3.3.6 Deep Learning Based Detection
3.4 Other Approaches
3.4.1 Unsupervised Learning and Anomaly Detection
3.4.2 Reinforcement Learning
3.5 Summary
References
4 Hybrid Model-Based and Data-Driven Detection
4.1 Concept Drift, Retraining, and Adaptiveness
4.2 Hybridization Approaches
4.2.1 Different Dimensions of Hybridization
4.2.2 Hybrid Model-Based and Data-Driven Ideas in Signal Processing and Communications
4.3 Feature Spaces Based onWell-Known Statistics or Raw Data
4.3.1 Nonparametric Learning: k-Nearest Neighbor
4.3.2 Quasi-Whitened Raw Data as Feature Vector
4.3.3 Well-Known CFAR Statistics as a Feature Vector
4.4 Rethinking Model-Based Detection in a CFAR Feature Space
4.4.1 Maximal Invariant Feature Space
4.4.2 Characterizing Model-Based Detectors in CFAR-FP
4.4.3 Design Strategies in the CFAR-FP
4.5 Summary
References
5 Theories, Interpretability,
and Other Open Issues
5.1 Challenges in Machine Learning
5.2 Theories for (Deep) Neural Networks
5.2.1 Network Structures and Unrolling
5.2.2 Information Theory, Coding, and Sparse Representation
5.2.3 Universal Mapping, Expressiveness, and Generalization
5.2.4 Overparametrized Interpolation, Reproducing Kernel Hilbert Spaces, and Double Descent
5.2.5 Mathematics of Deep Learning, Statistical Mechanics, and Signal
Processing
5.3 Open Issues
5.3.1 Adversarial Attacks
5.3.2 Stability, Efficiency, and Interpretability
5.3.3 Visualization
5.3.4 Sustainability, Marginal Return, and Patentability
5.4 Summary
References
List of Acronyms
List of Symbols
About the Author
Index
