Open AccessELM-Based Discriminant Auto-Encoder and Multi-Kernel Fusion for Radar Specific Emitter Identification
Author(s) -
Ya Shi,
Chenyi Wang,
Guangfu Ge,
Feixiang Liu,
Lele Zhang
Publication year - 2023
Publication title -
ieee access
Language(s) - English
Resource type - Journals
ISSN - 2169-3536
DOI - 10.1109/access.2023.3322375
Subject(s) - aerospace , bioengineering , communication, networking and broadcast technologies , components, circuits, devices and systems , computing and processing , engineered materials, dielectrics and plasmas , engineering profession , fields, waves and electromagnetics , general topics for engineers , geoscience , nuclear engineering , photonics and electrooptics , power, energy and industry applications , robotics and control systems , signal processing and analysis , transportation
Radar specific emitter identification (SEI) distinguishes different radar emitters, which is the research hotspot in the fields of electronic countermeasures and intelligence reconnaissance. To enhance the identification accuracy of the real radar SEI system with limited training data, we propose the multi-kernel extreme learning machine-based discriminant auto-encoder (MK-ELM-DAE) method by combining representation learning and multi-kernel fusion in this paper. Firstly, ELM-DAE is applied to each primary feature of radar signals to extract the more discriminative low-dimensional feature representations. With the features extracted by ELM-DAE, the linear discriminant ratio-based two stage multiple kernel ELM algorithm is then employed to conduct the multi-feature fusion. The most important module of MK-ELM-DAE is ELM-DAE, which is an effective supervised dimensionality reduction method for representation learning. ELM-DAE incorporates the label information into ELM auto-encoder by introducing a supersized regularization term, so it is more suitable for classification tasks. Specifically, we use the ambiguity function (AF) to extract primary features, and subsequently design an AF-based MK-ELM-DAE method for radar SEI. Experiments show that our method has significant advantages in accuracy and testing efficiency.