Artificial Neural Networks and Deep Learning Techniques Applied to Radar Target Detection: A Review

Radar target detection (RTD) is a fundamental but important process of the radar system, which is designed to differentiate and measure targets from a complex background. Deep learning methods have gained great attention currently and have turned out to be feasible solutions in radar signal processing. Compared with the conventional RTD methods, deep learning-based methods can extract features automatically and yield more accurate results. Applying deep learning to RTD is considered as a novel concept. In this paper, we review the applications of deep learning in the field of RTD and summarize the possible limitations. This work is timely due to the increasing number of research works published in recent years. We hope that this survey will provide guidelines for future studies and applications of deep learning in RTD and related areas of radar signal processing.

Deep Learning Network for Classifying Target of Same Shape using RCS Time Series

The main intension of this work is to find the warhead and decoy classification and identification. Classification of radar target is one of the utmost imperatives and hardest practical problems in finding out the missile. Detection of target in the pool of decoys and debris is one of the major radas technologies widely used in practice. In this study we mainly focus on the radar target recognition in different shapes like cone, cylinder and sphere based on radar cross section (RCS). RCS is a critical element of the radar signature that is used in this work to identify the target. The concept is to focus on new technique of ML for analyzing the input data and to attain a better accuracy. Machine learning has had a significant impact on the entire industry as a result of its high computational competency for target prediction with precise data analysis. We investigated various machine learning classifiers methods to categorize available radar target data. This chapter summarizes conventional and deep learning technique used for classification of radar target.

SOFTWARE-DEFINED RADAR TARGET SIMULATOR WITH DRFM FOR AUTOMATED RADARS TESTING

The paper describes a software-defined radar target simulator with DRFM. One may use such a device to test radars under development and evaluate their reliability in environments with radio frequency interference. The paper describes in detail a general algorithm of the target simulation and its modules for the simulation of range and speed. The authors also implemented the frequency shift module for the simulator in Xilinx System Generator and wrote this module in VHDL. One may find in the paper a comparison of the FPGA resources required for such implementations.

The five-point non-coherent model of a distributed radar object providing the independent control of angle noise probability characteristics along two orthogonal coordinate axes

The two-dimensional five-point non-coherent model replacing a distributed radar target is explored in this work. Four fixed model points are set in corners of the square but the fifth movable point lies inside of this square. Model points are supplied by normal uncorrelated random processes. The possibilities of the five-point non-coherent model of a distributed radar object for independent control of the producing angle noise parameters along two orthogonal coordinate axes are explored. The disadvantage of this model is noted - the connection of parameters values of angle noise probability density function for two coordinate axes. The expression describing this connection is specified. Expressions determining the boundaries of the allowable coordinate values of the fifth movable point of the five-point non-coherent model, within which the model provides the set parameters of the angle noise probability density function, are defined. The arrived results are validated by program simulations.

Method for long-term coherent-noncoherent signal accumulation with non-zero higher derivatives range to radar target

A method of long-term combined accumulation of the reflected signal is justified, which provides for its division into disjoint subsets, coherent accumulation in subsets using one of the fast algorithms and subsequent incoherent accumulation of the squares of the modules of the results of processing the subsets. A distinctive method’s feature is the use with incoherent accumulation of maxima of the squares of the moduli of the coherent processing results, that are selected from the range / radial velocity regions in accordance with a given hypothesis about the minimum and maximum values of the target radial velocity and the radial acceleration detection channel setting.The efficiency of the method was confirmed by simulation modeling. Using the theories of ordinal statistics and the method of moments, a method for calculating the probability of correct detection is developed. Estimates of processing losses are made in comparison with coherent and incoherent accumulation algorithms for a signal reflected from a point target, for the case when there is no range and frequency migration. Estimates for the required number of receiver channels are given.

Two-Stream Deep Fusion Network Based on VAE and CNN for Synthetic Aperture Radar Target Recognition

Usually radar target recognition methods only use a single type of high-resolution radar signal, e.g., high-resolution range profile (HRRP) or synthetic aperture radar (SAR) images. In fact, in the SAR imaging procedure, we can simultaneously obtain both the HRRP data and the corresponding SAR image, as the information contained within them is not exactly the same. Although the information contained in the HRRP data and the SAR image are not exactly the same, both are important for radar target recognition. Therefore, in this paper, we propose a novel end-to-end two stream fusion network to make full use of the different characteristics obtained from modeling HRRP data and SAR images, respectively, for SAR target recognition. The proposed fusion network contains two separated streams in the feature extraction stage, one of which takes advantage of a variational auto-encoder (VAE) network to acquire the latent probabilistic distribution characteristic from the HRRP data, and the other uses a lightweight convolutional neural network, LightNet, to extract the 2D visual structure characteristics based on SAR images. Following the feature extraction stage, a fusion module is utilized to integrate the latent probabilistic distribution characteristic and the structure characteristic for the reflecting target information more comprehensively and sufficiently. The main contribution of the proposed method consists of two parts: (1) different characteristics from the HRRP data and the SAR image can be used effectively for SAR target recognition, and (2) an attention weight vector is used in the fusion module to adaptively integrate the different characteristics from the two sub-networks. The experimental results of our method on the HRRP data and SAR images of the MSTAR and civilian vehicle datasets obtained improvements of at least 0.96 and 2.16%, respectively, on recognition rates, compared with current SAR target recognition methods.