Due to a serious shortage of training data, the performance of adaptive clutter suppression suffers remarkable degradation in heterogeneous environments. To address this problem, a novel clutter suppression method via affine transformation on manifolds is proposed. First, training samples in heterogeneous environments are characterized on an established manifold in which the distribution properties are analyzed. Then, a clutter classification scheme is proposed, whereby the KL divergence decision rule is derived to identify the training data as either homogenous or heterogeneous samples. Afterward, based on the distribution properties of samples and the clutter classification scheme, an affine transformation on manifolds is proposed for sample augmentation by transporting heterogeneous samples into the region of homogeneous samples. Finally, the clutter in the area of interest is suppressed on the manifold, which combines the transformed samples with the homogeneous samples, such that superior performance is obtained. Experiments on both simulated and real data validate the superiority of the proposed method in highly heterogeneous environments.

To improve the jamming cognitive level of radar in the complex and changeable electromagnetic environment, this paper proposes a YOLO-CJ (YOLO network for compound jamming detection) lightweight network for compound jamming signal detection. Firstly, the compound jamming image dataset is constructed by signal 0000-0000 © 2022 IEEE model and short-time Fourier transform (STFT). Secondly, based on the YOLOv7-tiny lightweight baseline model, the designed SimSPPFCSPC module, Triplet attention mechanism, and the C2f module are illustrated to improve the YOLOv7-tiny and establish the YOLO-CJ lightweight network, which aims to enhance the saliency and difference of the extracted features in the case of low SNR/JNR while effectively control the quantity of computation and parameters. Subsequently, combined with the compound jamming image dataset, the ablation experiment, algorithm comparison experiment, and robustness verification are carried out. The results demonstrate that the detection precision (mAP) and the inference time of a single image of the YOLO-CJ model can reach 98.69% and 12.73 ms, and the trade-off value of detection precision and speed can reach the highest 79.93%. Moreover, the mAP of the YOLO-CJ for compound jamming detection under SNR / JNR ≥ 0dB can reach more than 96%, which has good robustness and generalization ability under low SNR and JNR. Source code is released at https://github.com/zhuxuan-96/YOLO-CJ.

To treat the slow-moving fluctuating target detection issues, the information geometry-based track-before-detect algorithms, which combines the performance advantages of information geometry detectors and the multi-frame information integration ability of dynamic programming algorithm, are proposed in this paper. The proposed methods are based on the Hermitian positive definite (HPD) manifold and power spectrum (PS) manifold, that the geometric properties of the target feature track and clutter track on the HPD or PS manifold are exploited for the integration of inter-frame target information. Specifically, the IG-based scoring function is designed to evaluate the candidate trajectories, and a general IG-based DP algorithm is deduced to obtain the trajectories with the maximum integration of scoring functions. Moreover, to avoid the performance degradation resulting from the illegal candidate trajectories, a stricter constraint for the trajectories of slow-moving target is established, and a corresponding efficient IG-based DP algorithm is deduced by utilizing the skipping transition. Experimentally, the superiority of the proposed methods have been validated with the real-recorded ship target echo data and sea clutter data. The experimental results show that the proposed methods can provide a signal-to-clutter ratio improvement of approximately 5.5 dB.

Recently, the matrix information geometry (MIG) detectors have been speedily developed and demonstrated satisfactory performance in numerous applications. However, the MIG detector with a specific geometric measure is not always satisfactory under changing environments. This paper proposes the adaptive MIG detector with local metric tensor to match the detection environment for improving the adaptability of MIG detectors. The determination of the adaptive local metric tensor is formulated as an optimization problem based on the Neyman-Pearson criterion. Although this optimization is non-convex and constrained, this paper shows the equivalence between it and its Lagrange dual, and then a two-stage gradient descent method is designed to solve it. Moreover, the performance guarantee of the proposed method is provided, which reveals that it achieves asymptotically optimal detection performance among all MIG detectors. Experimentally, the adaptability and the effectiveness of the adaptive MIG detector are validated in three detection scenarios.

From the perspective of feature extraction and classification, deep neural networks are widely used in radar target detection. However, in heterogeneous environment, the traditional deep neural networks are difficult to extract a robust feature, which leads to the degradation of network detection performance. In order to address this problem, a radar target detection method with multi-task learning in heterogeneous environment is proposed. Considering the influence of heterogeneous data distribution, the proposed method designs a contrastive learning module added to a multi-task autoencoder. It can learn a compact and distinguishable feature representation, which enhances the feature separability between the clutter and the target. Simultaneously, a classifier is introduced to realize a binary detection in the feature representation. Comprehensive experiments are carried out to show that the proposed method guarantees a good detection performance in heterogeneous environment and solves the issue of over-fitting to a certain extent. Compared with some classical detectors, the proposed method shows better performance.

In this article, the power spectrum information geometry (PSIG) detector, which inherits the performance advantages of matrix information geometry (MIG) detectors in heterogeneous clutter backgrounds, is proposed. The PSIG detector can address two urgent problems in applications of MIG detectors, which are the expensive computation expense and unavailable acquisition ability of target velocity. Specifically, the PSIG detector utilizes power spectra (PSs) instead of high-dimensional covariance matrices to characterize sample data and employs subband filter bank to extend the detection from range cells to range-Doppler cells; thus, it requires less computation expense and can obtain the target velocity information according to the Doppler cell. Experiments based on the real data show the advantages of the proposed PSIG detectors in comparison with competitive methods. Especially, in the experiments with the real-recorded airborne radar data, the proposed method can effectively suppress the heterogeneous main-lobe clutter without any prior knowledge and provides detection probability improvement of more than 30% to the competitive methods with low false-alarm ratios.