Classification of ground moving targets using bicepstrum-based features extracted from Micro-Doppler radar signatures

Abstract

In this article, a novel bicepstrum-based approach is suggested for ground moving radar target classification. Distinctive classification features were extracted from short-time backscattering bispectrum estimates of the micro-Doppler signature. Real radar data were obtained using surveillance Doppler microwave radar operating at 34 GHz. Classifier performance was studied in detail using the Gaussian Mixture Mode and Maximum Likelihood decision making rule, and the results were verified on a multilayer perceptron and Support Vector Machine. Experimental real radar measurements demonstrated that it is quite feasible to discern three classes of humans (single, two and three persons) walking in a vegetation cluttered environment using proposed bicepstrum-based classification features. Sophisticated bispectrum-based signal processing provides the extraction of new classification features in the form of phase relationships in the radar data. It provides a novel insight into moving radar target classification compared to the commonly used energy-based strategy.

Operational assessment and adaptive selection of micro‐Doppler features

“…Over the past few decades a plethora of features have been proposed for use in micro‐Doppler classification. These features may be divided into four basic types: (i) physical features [7–9], which aim at deriving quantities relating to the physical characteristics of targets and their motion; (ii) transform‐based features [10–12], which utilise the coefficients of transforms, such as the discrete cosine transform (DCT), as features; (iv) component analysis features [13–15], where the basis computed from algorithms such as principle component analysis (PCA) are defined as features; and (iv) speech features [16–18], which have typically been designed for and used to process speech signals, but have yielded good results in micro‐Doppler classification as well. Combined, hundreds of features may potentially be extracted from micro‐Doppler signatures.…”

Micro-Doppler Based Classification of Human Aquatic Activities via Transfer Learning of Convolutional Neural Networks

“…In addition, another fruitful future research direction is a comparison of the performance of a DCNN and all combinations of existing feature-based schemes [6,7,8,9,10,11,12] to investigate the optimum performing methods for micro-Doppler signature-based human activity classification. …”

“…In particular, the unique micro-Doppler signatures from human activities enabled diverse and extensive research on human detection and activity classification/analysis using radar sensors [3,4,5,6,7,8,9,10,11,12]. More specifically, the authors of [6] extracted direct micro-Doppler features such as bandwidth and Doppler period, the authors of [7] applied linear predictive code coefficients, and the authors of [8] applied minimum divergence approaches for robust classification under a low signal-to-noise ratio environment.…”

“…More specifically, the authors of [6] extracted direct micro-Doppler features such as bandwidth and Doppler period, the authors of [7] applied linear predictive code coefficients, and the authors of [8] applied minimum divergence approaches for robust classification under a low signal-to-noise ratio environment. Furthermore, the authors of [9] suggested to use particle filters to extract features, the authors of [10] employed biceptrum-based features, the authors of [11] utilized orthogonal pseudo-Zernike polynomials, and features based on the centroid or the singular value decomposition (SVD) have been exploited in [12]. Compared to optical sensors, the electromagnetic radar sensors can operate in all weather conditions, regardless of lighting changes, and hence are competitive for the applications that require robust operation.…”

Learned Micro-Doppler Representations for Targets Classification Based on Spectrogram Images

“…Similarly, [18] applied Log-Gabor filters to extract features from time-frequency domain signals. Researchers also used features based on higher-order spectral processing extracted from Doppler echo signals [19], [20].…”