The digital radio mondiale (DRM)-based passive radar (PR) has now attracted wide attention for long-range aircraft detection. Use of long coherent integration time is an efficient way to compensate the large attenuation when the PR works via skywave propagation, but it also introduces nonstationarity into the echo signals due to target maneuver and ionospheric channel variation. This nonstationarity will severely degrade the performance of conventional constant false alarm rate (CFAR) detectors. To account for this problem, a novel processing scheme is proposed to improve the nonstationary target detection on the time-frequency (TF) plane. First, the TF multisqueezing transform (TFMST) is used to catch the TF signature and address energy divergence in the TF representation due to the ionospheric propagation. Next, morphological methodologies are applied to determine the regions containing potential targets. Finally, for each potential target, the duration, velocity, and energy are checked to decide whether it is a target. Simulation results show that, the detection probability of the proposed method is significantly increased compared with that of the existing methods for a low signal-to-clutter-and-noise ratio (SCNR). When tested on a dataset collected in a PR experiment with a 1150 km far DRM illuminator, the proposed method is able to detect targets up to 125 km that cannot be detected by those existing methods, which further proves its superiority for nonstationary target detection.

A high-frequency (HF) surface-wave radar has been widely used in marine monitoring and vessel surveillance, among which compact ones, e.g., a crossed-loop/monopole antenna (CMA), are preferred in many scenarios for their low costs and easy maintenance. Relatively large error in the direction of arrival (DOA) estimation due to the small aperture involved is one major problem in such compact radars. To solve this problem, a biologically inspired DOA estimation method is proposed to improve the direction finding (DF) performance. The phase amplification mechanism of Ormia ochracea is imitated with a digital signal processor appended to the radar receiver. Then, the multiple signal classification (MUSIC) algorithm is used to estimate the DOA. For a two-element array with a small spacing, simulation results show that the proposed Bio-MUSIC method is superior to the conventional MUSIC in DOA estimation, and the combination of cosine and sine loops offers significant advantages over that by using two monopoles in terms of both accuracy and stability over different look angles. Also, compared with the CMA, the root-mean-square error (RMSE) is reduced by up to 56.99%. Experimental results demonstrate that, the total RMSEs of the ship coordinates obtained by Bio-MUSIC based on two cosine loops or one cosine loop plus one sine loop are reduced by up to 41.23% and are much less than the RMSEs obtained by MUSIC based on the CMA in most cases. Consequently, the biomimetic antenna array (BMAA) of loops with Bio-MUSIC can also be a good choice for marine applications.

Compact high-frequency (HF) radar is a promising tool for maritime surveillance owing to its over-the-horizon and all-weather ability. However, it typically suffers from low tracking accuracy due to limited azimuth resolution. Cross localization using multiple radars provides a solution, but track-to-track association among radars remains a critical challenge. Previous association methods are mainly subject to that each radar’s range and azimuth measurement errors are comparable and sufficiently small, which is not satisfied in HF radar. In this article, we propose a novel track association algorithm based on the target region match (TRM) strategy, which employs the joint range and azimuth gates intersection criterion to combine the effect of range and azimuth measurement errors. Considering the azimuth error is much larger than that of the range, their contribution to the track association is balanced through the area intersection operation. Both simulation and experiment results show that the proposed TRM method produces better association performance than the other investigated methods.

Nearshore wind speed information is important for many research and commercial applications. Shore-based ultrahigh-frequency (UHF) radars promise all-weather, real-time, high temporal and spatial resolution measurements of nearshore wind. Until now, there are no reports on the method of measuring wind speed using the shore-based UHF radar. In this article, a method of wind speed extraction from the Bragg peak width of Doppler spectrum is proposed. Two models describing the relationship between the Bragg peak width and wind speed and direction are developed through theoretical derivation and numerical simulation. One is a combination of logarithmic functions and the other is a combination of power and logarithmic functions. The constants in these models are determined through data fitting. In the inversion process, the wind direction is first estimated, and then the wind direction and the Bragg peak width are substituted into the models to find the wind speed. Using the anemometer data of a buoy as the ground truth, it is proved that the method is feasible and the second model has better performance with a RMSE of 1m/s.

In addition to sea state monitoring, high-frequency surface wave radar (HFSWR) has attracted more and more attention in ship detection and tracking. Constant false alarm rate (CFAR) detectors have been widely used to handle the complex properties of sea clutters. Due to the relatively high detection threshold, the performance of CFAR in detecting weak and nonstationary targets is often not good. To address this problem, time–frequency analysis (TFA) is involved in the ship detection and direction finding (DF) processing. First, the probability distribution model of sea clutter is achieved in the time–frequency (TF) domain. Corresponding to this sea clutter model, the decision thresholds of targets on the TF plane under different false alarm rates$P_{\mathrm {fa}}$are calculated. Next, the array snapshots are formed by the spectral samples along each extracted TF ridge and later used in the DF process to give the direction of arrival (DOA). Experimental results show that, with the automatic identification system (AIS) records as the ground truth, the number of matched targets detected by the proposed TF-CFAR method is 5%–8% greater than that by the cell averaging (CA)-CFAR method. Moreover, the TF multiple-signal classification (MUSIC) also outperforms MUSIC with an improvement of 3.52° in the root-mean-square error (RMSE) of the DOA estimates under a low signal-to-noise ratio (SNR). In conclusion, the involvement of TFA can greatly improve the detection and DF performances of compact HF radar, particularly under the situations of low SNR and target nonstationarity.

Shore-based Ultra-high-frequency (UHF) radars have similar characteristics as the High-frequency (HF) ocean radars, for example, the sea waves generating Bragg resonance are gravity waves. Compared to HF radars, UHF radars are more sensitive to the small changes of sea waves, have finer spatial resolution and measure sea areas closer to the coast, hence are expected to complement the nearshore sensing capacity of HF radars. Two wind direction inversion methods for coastal UHF radars are adopted. One is based on the Bragg peak power ratio and the other is based on the modified Doppler spectrum centroid. The latter is proposed to address the problem that it is often difficult to calculate the Bragg peak power ratio at high wind speeds. In both methods, several models describing the relationship between the Doppler spectrum characteristics and wind direction are used respectively in combination with the least square method to estimate the wind direction with the ambiguity removed. Using the buoy wind sensor data as the ground truth, the results of a three-week experiment demonstrate the feasibility of the two methods. It is also shown that the former method has slightly smaller errors and the latter method applies to a wider range of wind speeds. For both methods, the most recommended model is the one based on the cosine-type directional spreading function with variable spreading parameter, because it has small errors and does not require data fitting.

Ship detection at the sea surface is important for improving human marine activities. Most existing ship detection methods for high-frequency surface wave radar (HFSWR) are based on peak and constant false alarm rate (CFAR) detection and require a coherent integration time (CIT) of several minutes. However, in such a long period, the target may not be stationary. To account for the nonstationary property, a time-frequency analysis (TFA)-based ship detection and direction finding (DF) method is proposed for HFSWR. Target ridges on the TF representation (TFR) of the echo data are detected first. Next, array snapshots are formed by sampling the extracted ridges and are used to estimate the direction of arrival (DOA). The processing results of the radar data collected at Dongshan, Fujian Province, China, show that the proposed method outperforms the CFAR method with both increased detection rates and decreased DF errors, especially under relatively low signal-to-noise ratio (SNR) scenarios.

The direction-finding high-frequency radars (HFRs) are widely used for remotely sensing the ocean surface currents. However, the performance of the direction-finding HFR is limited by the inevitable distortion of the actual antenna pattern of the compact monopole-cross-loop antenna, because the distortion of the antenna pattern leads to errors on the direction-of-arrival (DOA) estimation, which results in correctly determined radial velocities being placed into the incorrect bearing sectors. In this letter, a model determining the DOA estimation error (DOAEE) stemming from the antenna pattern distortion (APD) is presented with a detailed analytical derivation. This model suggests that the DOAEE only depends on the relative distortion of the two orthogonal loops. To validate the correctness of the proposed model, simulations are carried out. In addition, the simulation results show that the model well captures the actual relationship between the APD and the DOAEE. Moreover, a data set collected by a direction-finding HFR system and a synchronized automatic identification system receiver is used to further validate the proposed model. The precessing result of this data set demonstrates that the model-calculated DOAEE and the actual DOAEE are in good agreement with a correlation coefficient of 0.92 and a root-mean-square error of 9°, which provides solid evidence to verify the validity of the proposed model in determining the relationship between the DOAEE and the APD.

Wave-current interaction is a common and important phenomenon in the ocean. As an ocean remote sensing tool, high-frequency (HF) radar can be used to measure currents and wave parameters. In this paper, the possibility of studying wave-current interaction using HF radar is investigated. The first-order spectral power (FSP) of HF radar is used to explore the effect of current on the Bragg wave. By analyzing the FSP change with current (FSP-current distribution), we find that, in deep water, the wave-current interactions mainly belong to 2-D refraction case, while, over a relatively shallow shelf, the interactions are stronger and more complicated. Based on local topography and current field data at Taiwan Strait, the simulation results obtained using the SWAN model confirm the 2-D refraction of the Bragg wave. When the wave-current interaction is stable, we compensate the FSP with radar-measured currents according to the radar extracted FSP-current distribution and achieve a more accurate wind estimation. Comparisons between the original and refined wind fields show the effectiveness and necessity of the current-based compensation.

Wind spreading functions (WSF) are crucial for high-frequency radar (HFR) wind-direction inversion. The popular half-angle cosine WSF always fails to describe observed HFR Doppler spectra and tends to provide almost fixed relative angle estimations. In this letter, analysis of the data from a broad-beam HFR radar, deployed on the Taiwan Strait's west coast, shows that a modified WSF (based on the cosine WSF) has a better wind-direction estimation performance. The modified WSF fits average Bragg ratios of 15-day data well with the aid of data from buoys. The data of the next 13 days are used to test the modified WSF. The wind direction estimated by the modified WSF has an advantage when Bragg ratios have adequate average processing, and directions of arrival are around the upwind or downwind direction. The root mean square error of the modified WSF wind-direction estimate is 32.59° for the entire observation, decreasing to 14.18° when a significant wave height is between 1 and 2 m.

In addition to being able to extract current velocities and wind directions, the first-order Bragg peaks in high-frequency (HF) radar sea echoes also have the potential to independently map wind speeds. During a two-month experiment with the HF-radar ocean state measuring and analyzing radar, model S in the Taiwan Strait in 2013, the dependence of the first-order spectral power (FSP) on wind was found to be deterministic. An empirical model was thereafter proposed to estimate the wind speed, which gave a performance that was comparable with the popular second-order inversion method. Strong oscillations (up to ±3 dB) in the FSP caused by underlying currents were observed and cannot be readily explained by the classic wave-current interaction theory with the assumption of wave-action conservation. When we matched the local maximum and minimum values of the FSP with the local extreme and zero values of the radial-current velocities, we found that most of the extremes of the FSP occurred when the radial-current velocity was 0. To account for this phenomenon, additional smoothing with a 12.4-h moving window was applied, which led to a significant improvement in the wind-speed estimates. This knowledge of the FSP's wind and current dependence helps us make another step toward the operational estimation of wind speeds using the first-order Bragg peaks.

The conventional second-order Bragg-spectrum-based wave height extraction method is often susceptible to external noise and spatial aliasing. To improve the wave height estimate, we turn to the first-order Bragg peaks and propose a new method which directly estimates the wave height from them. The key point is to confirm and use the unsaturated property of the first-order Bragg spectral power. The quantitative relation between the first-order Bragg peak power and the significant wave height can be established with the help of an in situ wave buoy, and consequently, the wave height is to be read out from the curve via the maximum Bragg peak power on one range cell. The first-order method is validated by a two-month-long data set collected by the OSMAR-S radar at 13 MHz. Compared with the second-order method, the improvement is obvious under low and moderate sea states. The new method opens the way for wider use of the first-order Bragg peaks in wave height extraction by high-frequency radars.

Waveheight measurements with high-frequency radar have been troubled by the limitation dependent upon radio frequency for decades. It is suggested that high frequencies are needed for measuring low waves, whereas low frequencies are suitable for high waves. In this letter, a new waveheight inversion method using dual frequency is presented. Measurements of the two frequencies are fused together, according to their Doppler spectrum properties, to improve the range and accuracy of waveheight estimation. Experimental results collected on the coast of the Taiwan Strait over 18 days are shown to prove the validity of the method.

High-frequency (HF) radars have achieved great success in the remote sensing of sea states. Most operational methods for wave extraction resort to the second-order Doppler spectral continuum based on Barrick's formulas. However, accurate division of the first-order Bragg peak and the second-order continuum is often unavailable due to spectral broadening and splitting when the radar beam is not sufficiently narrow. The second-order peaks, particularly the second-harmonic peaks (SHPs), can also introduce large errors to the wave height estimates. In this letter, we describe some properties of the SHP that are directly learned from the experimental observations of sea echoes by HF radars in both the 25- and 13-MHz bands. The SHPs exist throughout two radar experiments with comparable strengths to the maximum of the second-order continuum. They have similar spectral structures as the Bragg peaks in both the Doppler and spatial domains, showing that these peaks are representations of some particular sea waves. The power ratios of the SHPs to the Bragg peaks (RSBs) in decibels have strong correlations with the wave heights recorded by an in situ buoy or a known model, and the linear regression models between the RSB and the wave height can be determined by data fitting. These findings give new insights into wave height estimation from the spectral peaks only, which can provide supplemental information to the conventional results from the second-order continuum to improve both the accuracy and robustness of the wave extraction by HF radars.

Ionospheric clutter resulting from undesired radio signals reflected by the ionosphere severely deteriorates the detection performance of the high-frequency surface wave radar (HFSWR) in the mid- and low-latitude regions. To solve this problem, we propose a novel antenna configuration and method, i.e., using vertically placed, coaxial, multilayer crossed-loops as auxiliary antennas to cancel the clutter on the uppermost main crossed-loop (and/or monopole) antenna. The difference signals between the main and the auxiliary loops with the same normal direction are used as the reference signals in the adaptive filtering scheme to avoid cancellation of the wanted signals, i.e., the echoes horizontally coming from the sea surface. Both simulation and some preliminary experimental results are demonstrated to show the validity of the method.

High-frequency (HF) radars have gained much attention in recent years due to their remarkable capabilities in remote sensing of sea surface states. Radio frequency interference (RFI) should be suppressed before extraction of useful information from the radar echoes. Element instead of array-based interference suppression method is preferred for a small-aperture radar, and it is still more difficult to deal with the nonstationary interference. In this letter, a short-time range domain cancellation method for RFI suppression is proposed, whose advantage over the global one is twofolded: better suppression performance and lower computation cost. Processing results of real data collected by OSMAR-S, a portable HF radar, show the validness of the short-time method for nonstationary RFI suppression. The method can greatly improve the anti-interference capability and detection performance of the HF radar.

Radio frequency interference (RFI) is often a major problem in high-frequency (HF) radar operation, and until now, few existing methods worked very well to suppress dense RFI. In this letter, a theoretical analysis of the range-domain correlation of RFI in HF chirp radar is presented, and the relation between the correlation functions of RFI in range and fast time domain is given; based on this, a new method for RFI suppression is developed. The echo signals at the reserved range bins are used to construct an interference subspace, and then, the signals at the range bins of interest are projected onto its orthogonal subspace to suppress the RFI. When applied to radar signal processing, the method suppressed dense nighttime RFI effectively. It greatly improves the radar's detection performance in dense RFI.