It is often difficult to obtain sufficient training data for adaptive signal detection, which is required to calculate the unknown noise covariance matrix. Additionally, interference is frequently present, which complicates the detecting issue. We provide a two-step method, termed interference cancellation before detection (ICBD), to address the issue of signal detection in the unknown Gaussian noise and subspace interference. The first involves projecting the test and training data to the interference-orthogonal subspace in order to suppress the interference. Utilizing traditional adaptive detector design ideas is the next stage. Due to the smaller dimension of the projected data, the ICBD-based detectors can function with little training data. The ICBD has two additional benefits over traditional detectors. Lower computational burden and proper operation with interference being in the training data are two additional benefits of ICBD-based detectors over conventional ones. We also give the statistical properties of the ICBD-based detectors and demonstrate their equivalence with the traditional ones in the special case of a large amount of training data containing no interference.

This paper investigates the problem of detecting a multichannel signal embedded in subspace interference and Gaussian noise. The interference lies in a known subspace but with unknown coordinates, while the noise, in the general sense, consisting of thermal noise and clutter, has an unknown covariance matrix. To estimate the covariance matrix, it is assumed that there are sufficient signal-free training data. Two cases are considered, namely, the homogeneous environment (HE) and partially homogenous environment (PHE). The test and training data in the HE have the same noise covariance matrix, while the test and training data in the PHE share the same noise covariance matrix up to an unknown scaling factor. We derive the corresponding Wald tests both in the HE and PHE. Remarkably, these two detectors reject the clutter by using the sample covariance matrix, and then reject interference and integrate signal simultaneously through the oblique projection. Numerical examples show that the two proposed detectors can provide better detection performance than existing detectors in some situations.

It is a compelling task to detect sea-surface small targets in the background of strong sea clutter. Traditional detection methods usually suffer from poor detection performance and a high probability of false alarm (PFA). In this letter, a PFA-controllable and feature-based detection method is proposed based on an enhanced convolutional neural network (CNN). The time–frequency (TF) features of received signals are first extracted by the short-time Fourier transform (STFT) and converted into feature images. These feature images are then employed as the inputs of an enhanced CNN with a PFA control unit. The enhanced CNN takes full advantage of the subtle feature extraction ability of the asymmetric convolution and robust TF maps (TFMs). Finally, detection results are obtained according to the given PFA. The results on the IPIX dataset show that the probability of detection (PD) of the proposed method is about 0.864 when the observation time is 1.024 s and PFA is 10 ⁻³ . Compared with five typical detection methods, the proposed method achieves better detection performance. Besides, results also verify the stable PFA control ability of the proposed method. The source code is available at https://github.com/quqizhe-whu/STDN .

This paper investigates the statistical performance of three adaptive detectors in the presence of subspace interference and Gaussian noise. The interference is deterministic and lies in a known subspace but with unknown coordinates, while the noise is Gaussian distributed with unknown covariance matrix. For performance evaluation, we consider a more general case, namely, the case of subspace signal mismatch, where the actual signal does not completely lie in the presumed signal subspace. We derive the exact statistical distributions of the detectors, and then obtain analytical expressions for probabilities of detection and false alarm. The theoretical study reveals that the interference and signal mismatch affects the detection performance through two angles, which are the angle between the interference subspace and actual signal, and the angle between the actual signal and presumed signal subspace after they are both projected onto the interference-orthogonalized subspace. Numerical examples are provided to verify the theoretical results.