Iet Radar Sonar and Navigation最新文献

This paper focuses on the problem of mainlobe interference suppression in radar networks tracking multi-target scenarios with multiple jammers. A beam allocation method is proposed based on stepwise suppression of mainlobe and sidelobe interference. Although a single radar can effectively suppress sidelobe interference, mainlobe interference necessitates collaborative countermeasures among multiple radar nodes, indicating that the beam allocation method significantly impacts interference suppression performance. We derive the mathematical relationship between target tracking accuracy and beam allocation strategy to address this. Subsequently, we establish a beam allocation model, with the beam allocation matrix as the optimisation variable and target tracking accuracy as the optimisation objective. One intelligent algorithm is employed to solve this model. Simulation results demonstrate that the proposed method optimises the allocation of anti-jamming resources in the radar network, leading to efficient interference suppression and stable target tracking.

Range migration (RM), Doppler frequency migration (DFM) and Doppler ambiguity, arising from high-speed manoeuvre of target, render moving target detection particularly difficult. In this paper, an efficient method based on reversal decoupling transform (RDT), scaled Fourier transform (SCFT) and keystone transform (KT), that is, RDT-SCFT-KT, is proposed for effective long-time coherent integration (LTCI) of an accelerated manoeuvring target with Doppler ambiguity. More specifically, an efficient RDT is performed to tackle range curvature and some other unnecessary terms. After that, the scaled inverse Fourier transform (SCIFT) and SCFT are successively introduced to handle the linear RM and DFM, thereby realising two-dimensional (2-D) energy integration. Besides, velocity and acceleration with respect to target could be acquired by the aid of peak detection. With the compensation associated with the obtained parameters, a linear and nonsearching approach involving the KT is developed to efficiently attain the final energy focusing and the remaining target parameters (i.e., range and unambiguous velocity). Given the above, the proposed RDT-SCFT-KT method could achieve LTCI for accelerated manoeuvring target and exhibits low computational complexity without requiring parameter searching process. Meanwhile, it remains effective for Doppler ambiguity (including velocity ambiguity and Doppler spectrum ambiguity) and offers complete motion parameters. Besides, the final energy integration is a linear transform process, which facilitates multi-target processing. The efficacy of the RDT-SCFT-KT integration approach is substantiated through numerical experiments.

In advanced defence and security systems, multi-sensor fusion is widely used to improve the overall observation capability, and heterogeneous sensors are a typical deployment in multi-sensor systems. Track-to-track association (T2TA) of heterogeneous sensors is the precondition and foundation of heterogeneous sensor track fusion. However, problems such as ubiquitous systematic and random errors, inconsistent update periods and features caused by two heterogeneous sensors bring significant challenges to T2TA and existing methods have not solved the above problems adequately. To address these problems, we propose conditional idempotent association generation for heterogeneous track-to-track association (CIAG). In CIAG, a track state mapping module (TSMM) is constructed to unify asynchronous and heterogeneous tracks from heterogeneous sensors. The TSMM can also mitigate the effects of systematic and random errors. An idempotent association generation module (IAGM) is constructed to model tracks and association matrices jointly, and generate association matrices directly and precisely. Moreover, CIAG realises an end-to-end generation from the track tensor to the association matrix that can avoid long time consumption caused by traversal calculations of tracks. Comprehensive experiments demonstrate that CIAG can achieve the best association performance and has better association efficiency.

The rotating corner reflector is widely used in passive jamming methods because of its micro-Doppler modulation effect on the echo signal of synthetic aperture radar. Nevertheless, the traditional methods based on this have problems of single jamming effect and limited jamming range, which are gradually unable to adapt to the increasingly complex electromagnetic countermeasure environment. Therefore, it is of great significance to optimise and improve this to achieve more abundant jamming effects. With the development of material technology, the electromagnetic meta-materials can realise the intra-pulse modulation of radar signals. Thus, a novel jamming method based on the uniform acceleration or deceleration rotating corner reflector and the periodic pulse modulation electromagnetic meta-material is proposed in this paper. Through changing the modulation parameters of proposed jamming model, four different types of jamming effects can be achieved and regulated flexibly. The imaging characteristics of proposed jamming model and the regulation effects of jamming parameters are analysed, respectively. Then, the simulation experiments are performed using the original echo data of airborne SAR system and the simulation results verify the correctness of theoretical analysis. Finally, the prototype design of proposed jamming model and the practical feasibility analysis are given, respectively, which can provide support for the subsequent engineering implementation.

In a distributed sensor fusion architecture, using standard Kalman filter (naive fusion) can lead to degraded results as track correlations are ignored and conservative fusion strategies are employed as a sub-optimal alternative to the problem. Since, Gaussian mixtures provide a flexible means of modelling any density, therefore fusion strategies suitable for use with Gaussian mixtures are needed. While the generalised covariance intersection (CI) provides a means to fuse Gaussian mixtures, the procedure is cumbersome and requires evaluating a non-integer power of the mixture density. In this paper, the authors develop a pooling-based fusion strategy using the harmonic mean density (HMD) interpolation of local densities and show that the proposed method can handle both Gaussian and mixture densities without much changes to the framework. Mathematical properties of the proposed fusion strategy are studied and simulated on two-dimensional (2D) and three-dimensional (3D) manoeuvering target tracking scenarios. The simulations suggest that the proposed HMD fusion performs better than other conservative strategies in terms of root-mean-squared error while being consistent.

Automatic object detection onboard drones is essential for facilitating autonomous operations. The advent of deep learning techniques has significantly enhanced the efficacy of object detection and recognition systems. However, the implementation of deep networks in real-world operational settings for autonomous decision-making presents several challenges. A primary concern is the lack of transparency in deep learning algorithms, which renders their behaviour unreliable to both practitioners and the general public. Additionally, deep networks often require substantial computational resources, which may not be feasible for many compact portable platforms. This paper aims to address these challenges and promote the integration of deep object detectors in drone applications. We present a novel interpretative framework, DetDSHAP, designed to elucidate the predictions generated by the YOLOv5 detector. Furthermore, we propose utilising the contribution scores derived from our explanatory model as an innovative pruning technique for the YOLOv5 network, thereby achieving enhanced performance while minimising computational demands. Lastly, we provide performance evaluations of our approach demonstrating its efficiency across various datasets, including real data collected from drone-mounted cameras and established public benchmark datasets.

For the issue of configuration difficulty in maintaining linear formations based on the same orbital plane for distributed space-based coherent aperture radar (DSCAR), it is necessary to modify the linear formation model into an arc formation model. This article derives the steering vector and joint pattern expressions for DSCAR based on uniform arc formation, and designs a segmented inertial factor (IF) particle swarm optimization (PSO) to seek the optimal solution for non-uniform spacing and random yaw angle in non-periodic geometric distribution. Simulation analysis shows that the combination of non-uniform spacing and random yaw angle in non-periodic geometric formations can achieve lower peak side lobe level (PSLL) compared to single non-uniform spacing and single random yaw angle but with wider beamwidth spread. Additionally, the segmented IF PSO proposed in this article balances convergence more quickly in the early stage of the search process and improves convergence speed to approach the optimal value (OV) in later stage. Compared with other IF PSO, it has better convergence speed and accuracy.

This paper investigates a distributed shipborne high-frequency hybrid surface-surface wave radar (HFHSSWR) model that combines shared sky wave paths with distinct shipboard surface wave paths. This model improves target localisation accuracy and overcomes the limited aperture of a single shipboard array. A weighted least squares (WLS) positioning algorithm based on a Gaussian Markov random field (GMRF) is proposed for the model. The algorithm converts the geodetic coordinates of measurement stations to Cartesian coordinates, then estimates the initial target position using bistatic range (BR) and time difference of arrival (TDOA) measurements. An iterative refinement approach is employed to mitigate discrepancies between spherical and planar models, utilising ionospheric altitudes extrapolated through a GMRF for enhanced positioning accuracy. Finally, target coordinates are converted back to geodetic form. Simulations indicate that this approach achieves higher positioning accuracy than standard WLS positioning algorithm.

Doppler-tolerant waveforms are some of the most common radar waveforms used in practice. However, their deterministic and repetitive nature impedes control of mutual interference when multiple radars operate in close proximity. Noise radar technology may address this problem but is not Doppler tolerant. In this study, we design a machine learning radar detector capable of Doppler-tolerant performance with noise waveforms. The detector is implemented as a feedforward multilayer neural network. We show that the detector may be trained to operate with one-bit data. Further, to evaluate the proposed detector's performance, we derive a closed-form expression of the receiver operating characteristic (ROC) for one-bit detection of a Swerling 1 target using the square-law detector under the assumption of low signal-to-noise ratio (SNR). Numerical results demonstrate that the proposed machine learning detector, when suitably trained, is capable of operating with Doppler tolerance over a wide range of Doppler shifts.

To overcome the limitations of existing methods in processing continuous acoustic signals—particularly issues related to modal aliasing and the constraints of Pekeris waveguide applications—this study proposes a depth estimation approach for continuous acoustic targets using hydrophone linear arrays. A horizontal linear array, designed to meet the resolution requirements of the F–K transform, is deployed to receive continuous acoustic signals. Environmental parameters are incorporated to fit the sound speed profile, and modal time-delay differences are calculated based on normal mode propagation models. Temporal compensation is then applied to each modal component of the received signals across array elements. The corrected signal matrix undergoes a bidirectional F–K transform transformation into the frequency–wavenumber domain, allowing for clear separation of the normal modes of continuous signals. Frequency–wavenumber curves are characterised based on the sound speed profile, and binary mask filters are designed to extract modal energy. Finally, a depth estimation matching function is constructed to facilitate energy search and matching. Simulation results indicate that the proposed method achieves depth estimation errors of less than 5% for 10-s broadband acoustic signals under negative sound speed profiles and real shallow-sea waveguide conditions. The method demonstrates improved stability and applicability in variable sound speed environments, offering greater practical value for real-world shallow-sea scenarios.