Iet Radar Sonar and Navigation最新文献

The two-sided correlation transformation (TCT) algorithm is widely used to estimate the direction of arrival (DOA) for broadband signals. However, the traditional TCT algorithm requires DOA pre-estimation, which results in a high computational complexity and poor performance under low signal-to-noise ratio (SNR) conditions. To address these challenges, an improved TCT algorithm based on array manifold interpolation (AMI) is proposed in this paper, which utilised the AMI method to decompose the array manifold matrix and reconstruct the signal covariance matrix. It aims to obtain a DOA-independent focusing transformation matrix, thereby avoiding DOA pre-estimation. The simulation and lake experiment results are compared with the traditional TCT algorithms. It shows that the proposed algorithm can achieve higher DOA estimation accuracy and better angular resolution even in low SNR environments by fully utilising the information within the whole bandwidth of the target while reducing computational complexity.

The inverse synthetic aperture radar (ISAR) system exploits the movement of the target to form its high-resolution image. Further, the multi-polarisation acquisition in ISAR collects additional information on the target's scattering properties and surface characteristics that help to enhance the imaging capabilities of ISAR. In this study, we suggest a novel multi-polarisation ISAR configuration based on the circular transmit and linear receive (CTLR) combination, namely CTLR hybrid-pol ISAR, for the application of non-cooperative target detection and imaging. The CTLR hybrid-pol ISAR captures sufficient information about the targets to accurately characterise them, and simultaneously overcomes the drawbacks of full-polarimetry (full-pol) ISAR associated with the transmission of two pulses to obtain a single unit of polarimetric back-scattered information. Validation is performed using real ISAR data of a T-72 tank target, collected under the moving and stationary target acquisition and recognition (MSTAR) programme conducted by the Georgia Tech Research Institute. A comparative analysis based on SPAN, entropy, and polarimetric decomposition is carried out between the full-pol ISAR and CTLR hybrid-pol ISAR information. The results conclude that CTLR hybrid-pol ISAR maintains a similar level of information content compared to full-pol ISAR while overcoming its drawbacks.

Method of moments is one of the most useful approaches for radar cross-section (RCS) simulation, allowing, that is, the computation of the scattering of real objects from 3D models. However, it is limited by computer memory and computation time. In this paper, the authors explore the question of the balance between the possible acceptable level of 3D model simplification and the time benefit associated with a decrease in computational overhead due to the reduction of the model geometry complexity. A spatial volume-based RCS characterisation quality index is proposed to help determine the level of simplification to achieve a significant reduction in computation time while maintaining an acceptable level of similarity. The authors present the results of the calculations performed for perfectly conducted sphere 3D models with varying levels of geometry simplification for which a simple analytical solution exists. Furthermore, the results of the computations performed for a generic missile model set are shown. Possible areas of the application of the proposed approach are also considered.

Existing integration of radar detection and communication (IDAC) systems are in general based on multi-input multi-output multi-stations or single-base transceiver splitting. However, these methods are challenging to realise IDAC for integrated receive-transmit half-duplex (IRTHD) pulse radars, which are detection-centric and are based on self-transmission and self-reception systems. The majority of recent studies in the field of IDAC for IRTHD pulse radars have focused on utilising time-division approaches to avoid conflicts, thereby also creating competition for radar time resources. In this paper, a pointer scheduling algorithm based on Tabu search (PS-TS) is proposed for IRTHD pulse radars, which solves the challenge of simultaneous efficient detection and communication. Firstly, the study presents a model for radar device-to-device (D2D) opportunistic communication and proposes a framework for pulse interleaving based on pointer scheduling to realise IDAC. Secondly, the PS-TS algorithm employs a Tabu search strategy to maintain high-quality solutions to avoid local optima and introduces a tolerance factor to maximise the communication success rate (CSR) with the minimal time expenditure. Simulation results indicate that the PS-TS algorithm outperforms the genetic algorithm and particle swarm optimisation in terms of robustness, CSR, and computational efficiency, providing real-time scheduling for IDAC systems.

This research investigates the problem of track conformation for a search and evade challenge within the context of the radar resource management domain. To analyse how agent collaboration affects the ability of multiple radar agents in confirming an evasive target, a high-fidelity radar simulation was designed. The simulated radar environment implements a realistic noise and clutter distribution and uses a generalised likelihood ratio test to make detections. The challenge is implemented as a limited information game with a highly evasive target attempting to reach one of four objective points before the track is confirmed by collaborative confirmation agents. Multiple Gaussian heuristic methods are compared with a reinforcement learning approach as the action selection agent. Three game theory strategies were also implemented: non-collaborative best response, collaborative best response, and leader-follower consensus. The results explore the effect of false alarms on confirmation performance and the impact of collaborative game theory applied to the agents versus isolated performance.

In this paper, the authors investigate and develop a low-cost synthetic aperture radar (SAR) testbed and utilise it to capture and reconstruct high-resolution 3D mmWave SAR images. Despite much attention from both industrial and academic scholars, cost restraints and complexity have hindered the deployment of SAR systems, particularly in the mmWave domain. This paper outlines the major components and systems of the 3D SAR testbed. The authors build a testbed with low-cost, commercially available hardware components and include open-source software that is highly reconfigurable and simple to reproduce. The multiple-input multiple-output (MIMO) radar sensor uses a frequency-modulated continuous wave (FMCW) chirp at the V-band (40–75 GHz) and synthesises a rectilinear two-dimensional planar aperture. The authors presented several reconstructed 3D images imitating real-world scenarios using the testbed. A non-linear sampling technique that has significantly decreased the amount of time required for data acquisition was implemented. In addition, the authors employed multiple image quality metrics to quantify and compare the images produced by the testbed. The testbed can be used to test and validate new techniques and algorithms.

The ability to control sidelobes in a SAR image is critical to forming images that are useful for interpretation and exploitation. QinetiQ has developed the RIBI sensing system, which utilises a distributed coherent array of sensors to produce multistatic images. These systems require techniques from outside the traditional radar domain to utilise the theoretical resolution possible in synthesising a coherent aperture from multiple disparate collections. This paper develops previously published work on using compressive sensing techniques to suppress sidelobes in SAR images to develop a higher-fidelity measurement model. Using Cranfield University's GBSAR System a series, experimental measurements are conducted, and image estimation techniques are applied to this real data. It demonstrates an improvement in recovery performance over an isotropic measurement matrix, and discusses areas which require further development.

Data association technology can determine the corresponding relationships between target data (measurements or tracks) on different platforms, laying the foundation for target allocation and collaborative strikes. For targets with smaller volumes and similar appearances, such as drone swarms and warhead swarms, angle measurement provides important information for the use of passive optical sensors to associate target data. The classic data association technology for angle-only measurement relies on precise knowledge of the sensor position, hindering their application to drone swarms or other low-cost mobile platforms. This paper proposes a data association algorithm based on the minimum angular distance sum, whose calculation process does not involve the sensor position information, making angle-only measurement data association technology independent of the sensor position information. Both the simulation results and the experimental results validated the effectiveness of the proposed algorithm.

Identifying the intentions of aerial targets is crucial for air situation understanding and decision making. Deep learning, with its powerful feature learning and representation capability, has become a key means to achieve higher performance in aerial target intention recognition (ATIR). However, conventional supervised deep learning methods rely on abundant labelled samples for training, which are difficult to quickly obtain in practical scenarios, posing a significant challenge to the effectiveness of training deep learning models. To address this issue, this paper proposes a novel few-label ATIR method based on deep contrastive learning, which combines the advantages of self-supervised learning and semi-supervised learning. Specifically, leveraging unlabelled samples, we first employ strong and weak data augmentation views and the temporal contrasting module to capture temporally relevant features, whereas the contextual contrasting module is utilised to learn discriminative representations. Subsequently, the network is fine-tuned with a limited set of labelled samples to further refine the learnt representations. Experimental results on an ATIR dataset demonstrate that our method significantly outperforms other few-label classification baselines in terms of recognition accuracy and Macro F1 score when the proportion of labelled samples is as low as 1% and 5%.

Multi-sensor networks often encounter challenges such as inconsistent sampling times among local sensors and data loss during transmission. To address these issues, this paper employs a data loss compensation strategy to reconstruct missing data information. It designs the state estimation of local sensors utilising iterative state equations, leveraging multistep prediction techniques to estimate sensor states at unsampled points, thereby transforming the asynchronous sensor network system into a synchronous one. Furthermore, the projection theorem is applied to determine the fusion weights of local sensors, grounded on the principle of square-averaging significance. Ultimately, data information pertaining to the same target is fused through arithmetic averaging, guided by distance correlation. Simulation outcomes demonstrate that the proposed algorithm balances estimation accuracy with communication overhead, achieved by designing an optimal number of communication iterations.