IET Radar, Sonar & Navigation

With the intelligent development of electronic countermeasures, the more complex and intense countermeasure environment poses a severe challenge to the anti‐jamming ability of radar. Aiming at the scene where radar detects and tracks targets, this paper proposes an optimisation method of a radar intelligent game anti‐jamming strategy based on the inference and induction of jamming behaviour with an OODA (observation, orientation, decision and action) loop. Firstly, the strategy selection of the OODA loop decision process of the jammer is optimised so that the jammer can predict radar behaviour. Secondly, the radar is given the ability to perceive the jammer strategy and infer the decision‐making ability of the jammer, and the radar behaviour with inductive ability is selected to generate the cover pulse so that the radar can recover the detection in time under the disadvantage situation, and the OODA loop cycle speed is improved under the advantage situation. At the same time, the judgement link of the jammer OODA loop is destroyed and the jammer's decision is induced to meet the radar expectation, and the success probability of the radar game against the intelligent jammer is improved comprehensively.

The use of recurrent waveforms in over‐the‐horizon radar (OTHR) necessitates techniques for ambiguity resolution and manipulation. This paper provides a number of techniques for manipulating the shape of radar ambiguity functions. A new and simple characterisation of the radar ambiguity function is introduced in terms of twisted convolution. It is shown that the ambiguity function of any waveform can be transformed by any desired area preserving linear transformation of the delay‐Doppler plane. Furthermore, given the desired delay‐Doppler transformation, the corresponding waveform transformation can be explicitly constructed through the factorisation of 2×2 $2\times 2$ matrices. Among other applications of this theory, it is shown that the usual OTHR phase and frequency coding techniques used for range‐folded spread Doppler clutter mitigation, which induce an approximate Doppler shearing of delay‐Doppler plane, can be replaced by chirp modulating the recurrent waveform. These non‐recurrent chirped waveforms induce an exact Doppler shearing and lead to simpler and more robust signal processing of the returns.

Passive localisation and tracking of a radio emitter is of significant interest for both civilian and defence applications. Among the existing methods, received signal strength indicator (RSSI)‐based localisation is widely used due to its low cost and simplicity. However, most RSSI‐based techniques make simplifying assumptions, such as relying on basic path‐loss models and presuming the emitter has already been detected, overlooking complex environmental effects like shadowing caused by obstacles. These limitations hinder the practical application of RSSI‐based methods. In this paper, we propose an advanced RSSI‐based framework for joint detection and tracking (JDT) of a radio emitter, integrating a more accurate propagation model that accounts for both path‐loss and shadowing effects. The emitter is modelled as a Bernoulli random finite set (RFS) characterised by its existence probability (EP) and spatial probability density function (SPDF), addressing the challenges of emitter detection uncertainty. A key innovation of this paper is the development of a fully distributed JDT algorithm, which overcomes the computational and communication challenges associated with centralised tracking systems. The proposed algorithm leverages parallel consensus on likelihood and prediction (PCLP), allowing for scalable and efficient operation across sensor networks. Simulation results validate the proposed method's performance in real‐time emitter tracking.

To improve the autonomous navigation accuracy of the Mars probe, a navigation method for orbit around mars using an auxiliary satellite and absolute and relative position information of x‐ray pulsars/inter‐satellite ranging/landmark integrated navigation is proposed in this paper. In this method, the Mars probe and the auxiliary satellite simultaneously observe the same x‐ray pulsar, and the difference in pulse arrival time (TDOA) is calculated by comparing their observations. The states of both the spacecraft and the auxiliary satellite are estimated by integrating the prior known position of the auxiliary satellite. To address systematic errors that remain constant over short periods—such as those introduced by the spacecraft's measurement instruments and satellite systems—these constant errors are incorporated into the state model to improve estimation and prediction accuracy. Moreover, to further enhance navigation precision, the approach integrates x‐ray pulsar navigation, inter‐satellite ranging, and landmark‐based navigation thereby improving system robustness. This approach demonstrates a significant reduction in errors, such as pulsar ephemeris inaccuracies and satellite clock drift, compared to traditional pulsar‐based navigation methods. Simulation results confirm the effectiveness of the proposed method in enhancing navigation performance.

The paper introduces and discusses the concept and findings of experimental trials focused on the passive multistatic radar localisation of low‐Earth orbit (LEO) space objects using terrestrial illuminators and a LOFAR (LOw‐Frequency ARray) radio telescope. In the considered solution, commercial terrestrial digital radio transmitters operating in the VHF band served as illuminators of opportunity, whereas the LOFAR radio telescope was employed as a surveillance receiver. The extensive antenna array of the LOFAR radio telescope enables the detection of relatively weak echo signals reflected from space objects moving in LEO. Reference signals were captured using software‐defined radio receivers placed near the illuminators of opportunity. By combining bistatic measurement results from three pairs of illuminator‐receiver, the position of a space object was estimated in a Cartesian coordinate system. The experimental results validate the feasibility of determining the position of space objects using a passive radar system that employs antenna arrays, such as the one in the LOFAR radio telescope, along with commercial terrestrial transmitters as illuminators of opportunity. The results of the performed simulations confirmed the accuracy of the object position estimation achieved in real‐life experiments.

An accurate stochastic model is essential for achieving high‐accuracy positioning solutions in the global navigation satellite system (GNSS) precise point positioning (PPP)/ultra‐wide band (UWB) tightly coupled (TC) integration. Conventionally, a priori variances are used in the PPP/UWB TC integration to determine the weights of observations. However, a priori variances are difficult to obtain in complex environments since the stochastic characteristics of different observations depend heavily on environmental conditions. By contrast, the variance component estimation (VCE) method can provide a more accurate stochastic model by estimating the measurement uncertainties of different types of observations. Nevertheless, the VCE is susceptible to measurements' outliers and low redundancy in complex observation environments. To address these issues, a robust stochastic modelling approach for PPP/UWB TC integration is proposed by optimising the VCE with a robust estimation strategy and an adaptive moving window filter technique. Two kinematic experiments are conducted in signal‐obstructed environments to validate the stochastic modelling approach. Results demonstrate that the three‐dimensional (3D) positioning accuracy in the PPP/UWB TC integration is improved by over 47% after VCE optimisation. Compared to the a priori variance‐based stochastic model, the robust stochastic modelling approach improves the 3D positioning accuracy by over 27%.

In the field of electronic reconnaissance, deinterleaving techniques for radar signals are crucial. Although a large number of studies have been devoted to the classification of known radar signals by recurrent neural networks under closed set conditions, this task remains challenging in open set environments. To this end, this paper introduces a novel variational autoencoder (LVAEGRU) based on gated recurrent units that incorporates a probabilistic ladder structure. This model aims at capturing higher level abstract features through probabilistic ladder structure, thus avoiding information loss at intermediate levels. By forcing the latent representation to approximate different multivariate Gaussian distributions and combining this with reconstructing the loss information, the method performs well in open‐set deinterleaving tasks. Experimental results show that the method proposed in this paper exhibits excellent performance in open‐set scenarios compared to multiple baseline methods. image

Millimetre‐wave road traffic surveillance radars are used for vehicle detection and tracking. Under rainy conditions, rain attenuation and rain backscatter severely degrade the vehicle detection performance of radars. In this paper, based on a big database of rain clutter collected by two radars at different rainfall levels, rain clutter modelling and the assessment of the vehicle detection ability of radars under rainy conditions are investigated. As the first contribution, the clutter‐to‐noise ratio (CNR) and signal‐to‐clutter‐noise ratio (SCNR) of the road traffic surveillance radar under rainy conditions are derived as functions of radial distance, based on the rain attenuation and backscatter models. The derived formulas highly accord with the measured CNR change of radars under rainy conditions, which can be used to evaluate radar performance and optimise the radar operating mode. As the second contribution, a range‐varying compound‐Gaussian model (CGM) with a clutter map cell partition is introduced to model rain clutter with range‐varying statistics, and a best‐type selection method of amplitude distributions is proposed. Based on the analysis of a big database of rain clutter at different rainfall levels, the range‐varying CGMs with gamma and lognormal textures are recommended to model rain clutter of millimetre‐wave road traffic surveillance radars.

Trust and explainability in localisation systems can be greatly helped by estimating a calibrated uncertainty. In this work, we argue for the first time that for this, it is best to express uncertainty in the location estimate directly rather than indirectly in the ‘noisiness’ or ambiguity of the data sample. Therefore, in this work, through a robust classification‐based model, we not only identify the most probable place but also provide a measure of confidence or uncertainty associated with the prediction of the place itself—in contrast to existing approaches where uncertainty values are produced with the same dimension as the encoded feature. We specifically prove the utility of this new formulation on CosPlace, a state‐of‐the‐art Geolocalisation system. Uncertainty is learnt by transforming Cosplace into an uncertainty‐aware neural network. To validate the effectiveness of our approach, we conduct extensive experiments using the Oxford Radar RobotCar Dataset, where we find that the backbone features learnt in the uncertainty‐aware setting result in better place recognition performance than vanilla Cosplace. Furthermore, by using it as a score to reject putative localisation results, we show that our uncertainty is well‐calibrated to place recognition accuracy—more so than two existing systems in uncertainty‐aware radar place recognition.

Space–time adaptive processing (STAP) can effectively detect moving targets in the background of ground clutter, but the performance will drop sharply when the training samples are limited. In this paper, to improve the clutter suppression performance when the training samples are limited, the authors propose a novel STAP algorithm based on sparse Bayesian learning (SBL) using a hierarchical synthesis prior. Firstly, we construct a novel three‐level hierarchical synthesis prior (HSP) model, which promotes the sparsity more significantly than traditional priors used in SBL. Secondly, in the framework of type‐II maximum likelihood approach, a novel iterative update criterion for hyperparameters is derived. Thirdly, in order to reduce the computational burden, the authors design a novel local space–time dictionary to transform the full‐dimensional clutter spectrum recovery problem into a local clutter spectrum recovery problem. Numerical results with both simulated and measured data demonstrate the excellent performance and relatively high computational efficiency of the proposed method.

The recognition of radar emitters modulation in an open‐set scenario presents a challenging task, particularly when identifying unknown modulation. This paper proposes a dictionary similarity based method for low intercept probability radar signal open‐set modulation recognition (OMR), designed to address the unknown modulation in open‐set scenarios. First, deep features of the input 1‐D signal are extracted, and a random Fourier transform is applied to map the signal into a high‐dimensional space, thereby converting the nonlinear feature optimisation problem into a linear optimisation problem. Next, an inter‐class discreteness (ICD) module and an intra‐class similarity (ICS) module are designed. Based on the Hilbert‐Smith independence criterion, the correlation between features is quantified, and the quantitative values of ICD and ICS are used as loss functions to constrain the network's learning process. This approach effectively enhanced the representational power of the class dictionaries and significantly improved the model’s overall performance. Experimental results demonstrate that the proposed strategy successfully extracts high‐dimensional feature prototypes, achieving high accuracy in closed‐set recognition while effectively performing open‐set recognition tasks.

In recent years, the use of millimetre wave radio signals for speech recognition has rapidly developed. The absence of high‐frequency components resulting from the material vibration constraints of fully viewed indoor objects has undermined the recognition accuracy in this field. This paper presents a new solution to the Chinese digits speech recognition problem by reconstructing the high‐frequency harmonic and non‐harmonic components with the radio signals received by millimetre wave radar sensors. A time–frequency analysis was conducted to convert the phase variations extracted from the radar I/Q signals to spectrograms. An improved threshold strategy was used to enhance the harmonic components on the spectrogram. Subsequently, a CycleGAN‐based network was constructed to recover non‐harmonic components on the spectrograms. An evaluation experiment was performed with a 77‐GHz frequency modulated continuous wave radar sensor to use the induced vibrations of aluminium foils, glass, and anti‐static bags to recognise the speeches of standard Chinese digit numbers (0–9). The F1 score in the speech recognition experiment reached 96.6%, with a micro average accuracy exceeding 98.3%. These results show that the proposed method can improve recognition accuracy by generating finer signatures from radio signals.

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.

Existing synthetic aperture radar (SAR) adversarial attack algorithms primarily focus on the digital image domain, and constructing adversarial examples in real‐world scenarios presents significant and challenging hurdles. This study proposes the template‐based universal adversarial attack (TUAA) algorithm. Initially, a SAR interference template generator module is constructed to derive a universal adversarial perturbation in the image domain. The designed loss function guides the parameter updating of the generator, thereby improving the attack effectiveness and perturbation concealment. Subsequently, a SAR jamming signal generator module is developed, which swiftly generates the interference signal using the range convolutional and azimuth multiplication modulation jamming method. Consequently, the victim model can be effectively targeted by merely transmitting the jamming signal to the SAR receiver. Experimental results show that TUAA reduces the recognition rate of four typical DNN models to less than 15% under acceptable time costs and image deformation.

For weak signal detection with direction‐finding (DF), this article presents a new receiver design approach that combines our accumulatively increasing receiver sensitivity (AIRS) signal detection algorithm with the compressive‐sensing (CS)‐based DF‐array/algorithm. The former uses the concept of timeslot (TS)‐based signal threshold detection, whereas the latter employs a frequency‐independent array with randomly located elements, whose bandwidth (BW) largely determines the DF‐array BW. To estimate the direction of a signal, the AIRS algorithm generates the array steering vectors in each TS when the amplitude of any frequency bins exceeds the predetermined threshold of the TS. The aim of this paper is to demonstrate the ability of the new receiver to detect low probability of intercept radar signals with high DF accuracy, fine frequency resolution, and good time‐of‐arrival measurement resolution. To discriminate accurate emitter directions from many false estimations created by the DF‐array in very low signal‐to‐noise ratio environments, K‐means clustering was also applied. In a scenario, the frequency modulated signals from several 165‐mW X‐band radars were in the field of view of a 6‐element DF‐array. Simulation results show that the receiver can accurately estimate all the emitters' directions with root mean squared error of less than 1°, when the separation between the DF‐array and radars is about 100 km.