250716_DAIRS_IEEE_Radar_rev_final_rev.pdf

XXX-X-XXXX-XXXX-X/XX/$XX.00 ©20XX IEEE

Defense Applications of Innovative Remote Sensing Frank C. Robey

Strategic Technology Office & Defense Sciences Office

Defense Advanced Research Projects Agency

Arlington, VA, 22203 frank.robey@darpa.mil, Benjamin Epstein ECS Federal

Arlington, VA, 22203

Kelly Carlson inPursuit LLC

Arlington, VA, 22203

Samuel L. Earp S.L. Earp Inc

Arlington, VA, 22203

Abstract—This paper introduces the Defense Advanced

Research Projects Agency (DARPA) Defense Applications of Innovative Remote Sensing (DAIRS) program. DAIRS is exploring technology development that could significantly improve the capabilities of over-the-horizon radar. Three topics are addressed in this program: mitigation of Doppler-spread, ionospherically-propagated clutter, maritime vessel classification, and the use of ambient noise as the illuminator for passive high frequency radar. The use of ambient noise for understanding the environment is becoming routine in geophysics, oceanography and radio astronomy. This program seeks to apply similar passive remote sensing techniques to radar.

Keywords—passive radar, over-the-horizon radar, surface wave radar, spread clutter, equatorial, ionosphere, disruption opportunity.

I. INTRODUCTION

Over-the-horizon radars (OTHR) operate in the high frequency (HF) band. This band creates significant challenges, particularly for those radars operating in low latitudes where scatter from the dynamic ionosphere creates clutter that can be spread throughout the Doppler space. In 2024 the Defense Advanced Research Projects Agency (DARPA) initiated a program, the Defense Applications of Innovative Remote Sensing (DAIRS) program to address ionospheric spread clutter and to explore two other challenges inherent in operating in the HF band, namely maritime vessel classification and the application of passive remote sensing methods to OTHR. A seminal paper that spurred the interest in the use of indigenous acoustic noise for oceanography [1] referred to the omnipresent acoustic noise illumination as acoustic daylight. DAIRS intends to bring the concept of radio frequency (RF) daylight into the radar domain.

The DAIRS program was released to potential performers as a DARPA Defense Sciences Office (DSO) Disruption Opportunity (DO), which is intended to be a rapid contracting and acquisition approach for innovative exploration and scientific investigations. From the time the solicitation was published to the program kickoff meeting, 120 days passed rather than the more typical 6 months to a year that occurs in more traditional solicitation approaches.

II. IONOSPHERIC SPREAD-CLUTTER MITIGATION

A. Physics A key goal of the DAIRS program is to investigate improving the ability of a surface wave radar to reliably provide awareness of the environment to the limit of the Exclusive Economic Zone (EEZ). Expanding reliable surface wave surveillance awareness to the 370.4km (200nmi) EEZ limit necessitates the use of frequencies in the lower part of the HF band, that is, around 5MHz, to reduce propagation losses. Using frequencies below the frequency where the signals do not propagate through the ionosphere but are reflected, referred to as the 𝑓𝑓𝑚𝑚𝑚𝑚𝑚𝑚 which currently peaks around 14MHz in low and mid-northern latitudes, results in undesired multipath returns.

These multipath returns occur for detection ranges greater than the height where ionospheric electron density is high enough to refract or backscatter the incoming radar transmission; that is, beyond about 90km for E-region returns, or beyond about 250km for F-region multipath. Fig. 1 illustrates a bistatic surface wave radar where spread Doppler clutter via an undesired ionospheric multipath modulates either a direct signal return to the receiver, or it modulates a path to the surface, which is then propagated into the receiver. This spread clutter can effectively mask detection of desired returns. An example of severe spread clutter is shown in Fig. 2, where the spreading from the ionospheric return is smeared severely in range due to the inability to effectively pulse compress.

The superposition of direct and ionosphere-reflected object signatures can become challenging to separate, as illustrated in Distribution Statement A. Approved for public release: distribution is unlimited.

Figure 1. Image illustrating the unwanted propagation from the ionosphere for a surface wave rave radar.

Desired path

Unwanted paths

Turbulent Ionosphere

Transmit

Receive mailto:frank.robey@darpa.mil

Figure 3. Here we can observe the characteristic Bragg scattering vertical lines from the sea surface, along with direct vessel responses between the Bragg lines and diffuse reflections from the ionosphere at ranges beyond about 300km illustrating the spread-Doppler nature of ionospheric clutter.

B. Mitigation approaches Various approaches have been proposed to alleviate this impairment of reliable radar operation. These approaches have included spatial adaptive nulling of the backscatter path [2-9] and the use of polarization [10]. Adaptive nulling on receive is insufficient to reject the transmitter-ionosphere and surface-receiver multipath since the angle of arrival and polarization of these paths can be coincident. The surface wave path for the desired signals is vertical linear polarization since this is the only mode that propagates over long distances. The sky wave path that reflects from the ionosphere, even with linearly polarized illumination, is generally elliptically polarized, so adaptive polarization nulling appears promising. However, the use of projection nulling to suppress the unwanted polarization modes appears to be insufficient; polarization super-resolution techniques need to be explored to understand how well polarization is able to suppress the unwanted returns. In the DAIRS program we intend to combine these approaches with multiple-input, multiple-output (MIMO) operation [11], similar to techniques developed for skywave radar, and recent decomposition-based signal processing techniques.

III. PASSIVE AMBIENT-NOISE RADAR

The geophysics, oceanographic, and astronomical sciences communities have developed sensing methods that use indigenous environmental noise sources for remote sensing. The geophysics community has successfully used acoustic noise from wind and microtremors, along with geophone arrays to characterize ice structure in Antarctica [12-13]. Similar examples can be found in oceanography [14-15] and radio astronomy. While current passive radar methods are primarily based on estimating or knowing the temporal properties of the waveforms, this program will build on the body of work covering passive remote sensing to investigate and develop techniques for passive noise radar, without the need to know or estimate the temporal properties of the received waveforms.

Since beginning this program, an early study on the possibility of using passive techniques at HF has been brought to our attention [17-18].

One concept for passive radar using indigenous ambient noise is shown in Fig. 4. One or more receive arrays collect and support the processing of the signals received from the environment, which includes those signals that propagate directly from the noise field to the receiver and those that are scattered from objects in the scene and then propagate to the receive array.

A. Theoretical aspects We illustrate the potential for object detection assuming that we have two receiving apertures located within a noise field that is uniform, isotropic and with sources located at such distance that they can be modeled as being from infinite range. A simplified model of the two receivers, 𝑠𝑠1and 𝑠𝑠2, located at 𝑠𝑠1 = {𝑥𝑥, 𝑦𝑦, 𝑧𝑧} = �0,0, 𝑑𝑑

� and, 𝑠𝑠2 = {𝑥𝑥,𝑦𝑦, 𝑧𝑧} = �0,0,−𝑑𝑑

� , with a scattering object, 𝑣𝑣, and a single noise source, 𝑛𝑛𝑖𝑖 are shown in Figure 5. In this model, the signal received by each of the apertures are the sum of the direct path from the noise source to

Figure 2. Example of severe spread clutter from a 2018 measurement campaign in Palau. Courtesy ONR.

R ad ar R an ge km

Local Time of Day

1400 0200

Spread clutter

Ground wave

1 hop F

2 hop F

Day Night

Es

Po w er dB

Figure 4. Passive illumination using ambient emitters for radar.

Figure 3. Range-Doppler Map obtained from an NRL SWOTHR array at Wallops Island, VA. The color coding represents distinguishing features. Strong clutter returns originate via various ionospheric paths. (DARPA figure) the receiver plus the path that is scattered to the receiver by way of the object in the field. In the frequency domain, the received signals referenced to the origin at each of the apertures are 𝑠𝑠1(𝜔𝜔) = 𝑛𝑛𝑖𝑖(𝜔𝜔,𝜃𝜃,𝜙𝜙)𝑒𝑒 𝑗𝑗𝑗𝑗𝑗𝑗cos𝜃𝜃

2𝑐𝑐 + 𝑛𝑛𝑖𝑖(𝜔𝜔, 𝜃𝜃,𝜙𝜙)𝑒𝑒 𝑗𝑗𝑗𝑗(𝑛𝑛�𝑖𝑖⋅𝑣𝑣��⃗ ) 𝑐𝑐 𝑒𝑒− 𝑗𝑗𝑗𝑗𝑅𝑅1

𝑐𝑐 (1) 𝑠𝑠2(𝜔𝜔) = 𝑛𝑛𝑖𝑖(𝜔𝜔, 𝜃𝜃,𝜙𝜙)𝑒𝑒− 𝑗𝑗𝑗𝑗𝑗𝑗cos𝜃𝜃

2𝑐𝑐 + 𝑛𝑛𝑖𝑖(𝜔𝜔, 𝜃𝜃,𝜙𝜙)𝑒𝑒 𝑗𝑗𝑗𝑗�𝑛𝑛�𝑖𝑖⋅𝑣𝑣��⃗ � 𝑐𝑐 𝑒𝑒− 𝑗𝑗𝑗𝑗𝑅𝑅2

𝑐𝑐 (2)

Where 𝜔𝜔 is the center frequency, 𝑅𝑅1 and 𝑅𝑅2 are the distances from the object to the first and second receive aperture respectively, 𝑛𝑛�𝑖𝑖 is a unit vector in the direction of 𝑛𝑛𝑖𝑖, and 𝑣⃗𝑣 is a vector from the origin to the scattering object. The inner product, 𝑛𝑛�𝑖𝑖 ⋅ 𝑣⃗𝑣, indicates the projected distance in the direction of the noise source away from the origin. In this equation, the receive effective height, which converts from electric field strength to received signal, and the object’s scattering coefficient are assumed to be unity.

Multiple authors have derived the narrow-band correlation between two receivers with isotropic antennas in an isotropic noise field [17-20] when integrated over the sphere. As a result, we can readily write the correlation of the signals received by the two apertures, 𝐶𝐶12(𝜔𝜔) as

𝐶𝐶12(𝜔𝜔) = 𝐼𝐼(𝜔𝜔) �sinc �𝜔𝜔𝜔𝜔 𝑐𝑐 � + 𝜎𝜎

𝑅𝑅1

𝑒𝑒 𝑗𝑗𝑗𝑗𝑅𝑅1

𝑐𝑐 sinc �𝜔𝜔𝑅𝑅2 𝑐𝑐 𝜎𝜎

𝑅𝑅2

𝑒𝑒− 𝑗𝑗𝑗𝑗𝑅𝑅2

𝑐𝑐 sinc �𝜔𝜔𝑅𝑅1 𝑐𝑐 � + 𝜎𝜎2𝑒𝑒− 𝑗𝑗𝑗𝑗 𝑐𝑐 (𝑅𝑅1−𝑅𝑅2)

𝑅𝑅1𝑅𝑅2

� (3)

While (3) is based on the narrow-band assumption, it illustrates the potential for object detection and localization. The first term in the brackets represents the isotropic noise received directly by each of the apertures and describes the correlation of the signals received by each of the apertures as a function of the distance between the apertures. The second term is due to noise being directly received by the first aperture, (𝑠𝑠1 ), and through a scattered path created by the object into the second aperture. The third term is due to noise scattered by the object then received at the first aperture and the noise received directly by the second aperture. The fourth term inside the brackets is due to the noise impinging on the object that then scatters and propagates into each of the apertures. The term 𝜎𝜎 is a factor to account for the energy that is scattered by the object, and which undergoes voltage spreading into each of the apertures accounting for the 1/𝑅𝑅1 and 1/𝑅𝑅2 terms. In a more typical radar equation, 𝜎𝜎2 as shown in the fourth term in the brackets of this equation would be considered the radar cross section.

If we simplify (3) by assuming that the range to the object is large relative to the spacing between the apertures, 𝑑𝑑, and that the object is at an angle of 𝜃𝜃𝑣𝑣, then we can make the following approximation: 𝑅𝑅1 ≈ 𝑅𝑅 − 𝑑𝑑 cos 𝜃𝜃𝑣𝑣 , 𝑅𝑅2 ≈ 𝑅𝑅 + 𝑑𝑑 cos𝜃𝜃𝑣𝑣 , 1

𝑅𝑅1

≈ 1

𝑅𝑅

and 1

𝑅𝑅2

≈ 1

𝑅𝑅

. Then (3) can be rewritten explicitly:

𝐶𝐶12(𝜔𝜔) = 𝐼𝐼(𝜔𝜔) �sinc �𝜔𝜔𝜔𝜔 𝑐𝑐 𝜎𝜎

𝑅𝑅

�𝑒𝑒𝑗𝑗𝑗𝑗(𝑅𝑅−[𝑑𝑑/2]𝑐𝑐𝑐𝑐𝑐𝑐𝜃𝜃𝑣𝑣)/𝑐𝑐sinc � 𝜔𝜔�𝑅𝑅+𝑑𝑑2 cos𝜃𝜃𝑣𝑣� 𝑐𝑐 𝑒𝑒−𝑗𝑗𝑗𝑗(𝑅𝑅+[𝑑𝑑/2]𝑐𝑐𝑐𝑐𝑐𝑐𝜃𝜃𝑣𝑣)/𝑐𝑐sinc � 𝜔𝜔�𝑅𝑅−𝑑𝑑2 cos𝜃𝜃𝑣𝑣� 𝑐𝑐 �� + 𝜎𝜎2𝑒𝑒− 𝑗𝑗𝑗𝑗𝑗𝑗 𝑐𝑐 cos𝜃𝜃𝑣𝑣

𝑅𝑅2

� (4)

This equation illustrates how the range and angle of arrival of the object enter into the physics of the cross-correlation of the signals received by two apertures providing the potential to estimate range and bearing. What might not be clear from this equation is that the first and last terms within the brackets, after a Fourier transform, result in responses that are concentrated at relatively short time delay.

Multiple authors illustrate that the absolute value of the derivative of the Fourier transform of the first term in (4) is the forward and reverse Green’s function between the two apertures [12,13,15]. The sinc(. ) function in the inner terms can be expanded using the Euler relation for the sine function. The Fourier transform of these terms produces peaks at the values ±2𝑅𝑅 and −𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑(𝜃𝜃𝑣𝑣). The range and the angle of the object can be determined from these peaks.

The challenge that is exemplified by (4) is how to perform the signal processing to detect and estimate the parameters of the object. While this has similarity to previous work in generalized cross correlation for time delay estimation, there are enough differences that those techniques do not apply directly, and thus new algorithms must be developed.

B. Research There are many significant challenges in the development of radar that uses indigenous noise as the illumination. Some of these challenges are quite basic, while some are complex and may require extensive research. Examples of the challenges include the following.

• Coherence of the illumination between areas of interest and the direct path. Decorrelation may occur and detract from the ability to discern an object or cell of interest from clutter.

• Determination of what the appropriate radar range equation should be. It is apparent that the radio astronomy radiometry equation plays a factor, where the sensitivity increases as the square root of the time-bandwidth product.

• Management of observed area dispersion. Processing over sufficient bandwidths is needed to accentuate features of the cell of interest with the result that the

Figure 5. Coordinate system illustrating a single noise source.

Doppler frequency cannot be modeled as constant. The variation in Doppler frequency must be inherent in the estimation processing.

• Management of environmental dispersion. Processing of data in a maritime environment since the Bragg resonant frequency, 𝑓𝑓𝐵𝐵 of ocean surface gravity waves varies as the square root ratio of the earth’s gravitational constant and the wavelength, 𝜆𝜆, or 𝑓𝑓𝐵𝐵 = �𝐺𝐺𝑒𝑒 𝜋𝜋𝜋𝜋

(5)

• Anisotropic, colored, or multipath noise which potentially masks the features from the cell of interest.

The height that signals refract from the ionosphere is frequency dependent and at higher frequencies the signals propagate through rather than refract from the ionosphere. This phenomenon will result in both an anisotropy as well as dispersive relationship for the illumination.

The losses as the scattered signals propagate across the ocean’s surface as a surface wave is frequency dependent and can be anisotropic due to sea state or land masses.

• Anisotropic and/or frequency dependent scattering from the cell of interest.

• Polarization of the illuminating electromagnetic energy

vs. the platform polarimetric response and the channel response between the platform and the receiver.

• Sampling bit depth versus wider bandwidth. In radio astronomy there is a clear trade in favor of fewer bits in each sample, since reducing the number of bits increases the bandwidth that can be processed.

Development of signal processing to perform clutter rejection and moving object detection that mitigates these challenges to show the potential for passive ambient noise radar is at the heart of this early-stage DARPA research programs.

IV. MARITIME VESSEL CLASSIFICATION

Narrow bandwidth radars, particularly those operating in the

HF band, are typically not used to classify observed maritime ship traffic. The long wavelengths used for HF radar, particularly at the lower frequencies that support long range propagation, can be of vehicle scale, which has historically prohibited the identification of platform features that are typically used for classification purposes in microwave radars.

The HF surface wave radar bands lie within frequency allocations provided by the International Telecommunication Union (ITU) and the United States (U.S.) National Telecommunications and Information Administration (NTIA).

These bands are discontinuous and do not have what would be considered a single “wide band” allocation. As a result, wider bandwidth data collections will necessarily be discontinuous and straightforward Fourier-type processing must be enhanced with sparse frequency band processing when wider bandwidths are needed to provide resolution on the vessels.

Multiple authors have published papers describing approaches to maritime vessel classification using HF surface wave radar [23-29]. In general, these techniques have not yet produced reliable classification of maritime vessels. DAIRS is exploring the use of longer observation periods along with multi-frequency and multi-angle observation processing to classify maritime vehicles. The intent of this investigation is to perform a “first principles” look at classification for low frequency and relatively narrow band radar sensors. Figure 6 shows radar tracking of ships off the coast of Virginia, acquired with a Naval Research Laboratory (NRL) Surface Wave Over-the-horizon Radar (SWOTHR) array. Such data is matched to maritime Automated Information System (AIS) to serve as training data for classification developments in DAIRS. To augment the collected data in support of approaches that need large training and test data sets, radar cross section (RCS) models are being constructed based on known vessel types (e.g., as extracted from AIS data).

V. DEVELOPMENT AND TEST ACTIVITIES

The exploration of the three topic areas that make up the

DAIRS program is based on a firm physical foundation. The development is driven by a robust data collection and test campaign.

The first data sets consisted of combinations of several apertures. NRL operated their wideband transmit array, located at Cape May, New Jersey in all data collection events. Rutgers University operated their Coastal Dynamics Radars (CODARs) to provide bistatic illumination for some of the test events.

Figure 6. Footprints of the NRL and DAIRS arrays located on the east coast of the United States. The inset is an image of one of the ships tracked in the overlap region that is being used in vessel classification. Maersk Bahamas image from Vesselfinder.com. Background is from Google Earth.

Multiple receive arrays have been used for the data collection events held to date. The arrays include an NRL-built 2-D array of monopoles, and two linear and 2-D arrays of eight three-axis polarimetric antenna (see Figure 7). Additionally, Georgia Tech operated a lower frequency set of antenna and receivers oriented toward collecting lightning data.

The data collected at the time of this paper’s submission were in the United States Eastern Seaboard region between Cape May, New Jersey and Virginia Beach, Virginia. These data provided an excellent cross section of mid-latitude ionospheric conditions and included a variety of vessel classes as determined by the ships broadcast of their Automated Information System (AIS) information.

In the next phase of the program, the focus of the data collections will move to low latitudes in the mid Pacific. While the vessel traffic is lower, the intent is to collect data where the turbulence of the equatorial anomalies and equatorial spread-F clutter provides data that will be far more challenging for clutter mitigation (see Figure 2). The intent for the low latitude collections is to operate the transmitters at multiple frequencies and polarization in a MIMO radar mode while receiving via two geographically spaced polarimetric arrays. Ionospheric sounding and other instrumentation will enable characterization of the ionospheric conditions and provide insight into the physics that underlie the measurements and the data that will be collected.

VI. SUMMARY

The DARPA DAIRS program is an exploration of new approaches to provide reliable long-range detection of maritime traffic throughout the Exclusive Economic Zone. Performers for this program were solicited through a DSO Disruption Opportunity, which is intended to enable rapid investigation of high-risk science. To improve the quality and usefulness of surface wave radar returns, DAIRS aims to develop new approaches to spread-Doppler clutter mitigation and maritime object classification. The DAIRS program will be performing data collections and investigations into the use of ambient indigenous noise as radar illumination with a focus on the use for understanding the environment. Extensive collections of SWOTHR return data during DAIRS are central to the methods developed under the current and future SWOTHR programs.

These data will be used to develop technology in the three areas described in this paper. Addressing the challenges described in this paper are at the heart of this early-stage DARPA research program.

VII. ACKNOWLEDGMENTS

The support and data collection by the Naval Research

Laboratory for the development and operation of the sensors used to collect the data for this program is greatly appreciated.

The antenna arrays pictured in Figure 7 were designed, built and deployed by Dr. Alan Fenn and Dr. Brad Perry of MIT Lincoln Laboratory. The cooperation of Rutgers University, Dr. Hugh Roarty, to modify the operation of the Coastal Dynamics Radar (CODAR) in support of DAIRS data collections is appreciated.

The National Aeronautics and Space Administration allowed use of their facilities for the deployment of DAIRS arrays. The authors acknowledge the generosity of the Office of Naval Research for use of the example shown in Figure 2.

Figure 7. An array of 8 three-axis multi-polarization antennas.

Deployment was in support of the DAIRS program to provide spatial and polarization diverse data. DARPA photo.

VIII. REFERENCES

[1] M. Buckingham, B. Berknout, and S. Glegg, “Imaging the ocean with ambient noise.” Nature 356, 327–329 (1992).

https://doi.org/10.1038/356327a0

[2] G. S. Antonio and Y. I. Abramovich, "Two-dimensional High Frequency Surface Wave Radar receive array design," 2017 IEEE Radar Conference (RadarConf), Seattle, WA, USA, 2017, pp. 1402-1407, doi:

10.1109/RADAR.2017.7944426.

[3] T. Wenlong, L. Gaopeng and X. Rongqing, "Ionosphere interference suppression for HFSWR using L-shape array," 2010 2nd International Conference on Signal Processing Systems, Dalian, China, 2010, pp. V1- 484-V1-488, doi: 10.1109/ICSPS.2010.5555544., [4] H. W. H. Leong, "A comparison of sidelobe cancellation techniques using auxiliary horizontal and vertical antennas in HF surface wave radar," Record of the IEEE 2000 International Radar Conference [Cat. No.

00CH37037], Alexandria, VA, USA, 2000, pp. 672-677, doi:

10.1109/RADAR.2000.851914.

[5] G. A. Fabrizio, A. B. Gershman and M. D. Turley, "Non-stationary interference cancellation in HF surface wave radar," 2003 Proceedings of the International Conference on Radar (IEEE Cat. No.03EX695), Adelaide, SA, Australia, 2003, pp. 672-677, doi:

10.1109/RADAR.2003.1278823.

[6] A. M. Ponsford, R. M. Dizaji and R. McKerracher, "HF surface wave radar operation in adverse conditions," 2003 Proceedings of the International Conference on Radar (IEEE Cat. No.03EX695), Adelaide, SA, Australia, 2003, pp. 593-598, doi: 10.1109/RADAR.2003.1278808.

[7] S. Shang, Zhang Ning and Li Yang, "Extraction of ionospheric clutter in HFSWR," 2010 The 2nd International Conference on Industrial Mechatronics and Automation, Wuhan, China, 2010, pp. 296-299, doi:

10.1109/ICINDMA.2010.5538312.

[8] A. M. Ponsford, R. M. Dizaji and R. McKerracher, "HF surface wave radar operation in adverse conditions," 2003 Proceedings of the International Conference on Radar (IEEE Cat. No.03EX695), Adelaide, SA, Australia, 2003, pp. 593-598, doi: 10.1109/RADAR.2003.1278808.

[9] Ji, X., Wu, X., Yang, Q, Zhang, X..: Overview of ionosphere clutter suppression for high frequency surface wave radar (HFSWR) system:

Observation, approaches, challenges and open issue. IET Radar Sonar Navig. 17(12), 1743–1759 (2023). https://doi.org/10.1049/rsn2.12476

[10] Y. Yang, X. Mao, Y. Hou and J. Geng, "A Two-Step Method for Ionospheric Clutter Mitigation for HFSWR With Two-Dimensional Dual- Polarized Received Array," in IEEE Access, vol. 8, pp. 105903-105913, 2020, doi: 10.1109/ACCESS.2020.2999463.

[11] G. J. Frazer, Y. I. Abramovich, B. A. Johnson and F. C. Robey, "Recent results in MIMO over-the-horizon radar," 2008 IEEE Radar Conference, Rome, Italy, 2008, pp. 1-6, doi: 10.1109/RADAR.2008.4720867.

[12] A. Sergeant, M. Chmiel, F. Lindner, F. Walter, P. Roux, J. Chaput, F.

Gimbert, and A. Mordret, “On the green’s function emergence from interferometry of seismic wave fields generated in high-melt glaciers:

implications for passive imaging and monitoring,” The Cryosphere, vol.

14, no. 3, pp. 1139–1171, 2020. [Online]. Available:

https://tc.copernicus.org/articles/14/1139/2020/

[13] J. Chaput. (2008) Green’s functions, noise, and seismic interferometry.

Accessed: 2025-03-08. [Online]. Available:

https://www.scribd.com/document/241494510/chaput

[14] W. A. Kuperman. (2007, June) Extracting Green’s functions from random noise and vibrations. Ira Dyer Symposium, MIT. Accessed: 2025-03-08.

[Online]. Available: https://acoustics.mit.edu/dyerparty/dyer kuperman.pdf

[15] C. Bardos, J. Garnier, and G. Papanicolaou, “Identification of green’s functions singularities by cross correlation of noisy signals,” Inverse Problems, vol. 24, no. 1, 2008

[16] K. F. Woolfe, S. Lani, K. G. Sabra, and W. A. Kuperman, “Monitoring deep-ocean temperatures using acoustic ambient noise,” Geophysical Research Letters, vol. 42, no. 8, 2015.

[17] R. Shuchman, N. Subotic, J. Burns, and C. Roussi, “Exploitation of arctic phenomena for unattended sensing,” May 2013, unpublished Assured Arctic Awareness Program briefing.

[18] R. Shuchman, D. Atwood, J. Burns, B. Hart, C. Roussi, and B. Wilson, “Hf passive interferometric tracking for maritime surveillance,” Michigan Tech Research Institute, Ann Arbor, MI, Tech. Rep. MarineRF FR201501, 2015, unpublished.

[19] H. Cox, “Spatial correlation in arbitrary noise fields with applications to ambient sea noise,” Journal of the Acoustical Society of America, vol. 54,

no. 5, pp. 1289–1301, May 1973.1892, pp.68–73.

[20] H. Cox, H. Lai, and K. Bell, “Spatial correlation for directional sensors in arbitrary noise fields,” Proc. Asilomar Conf on Signals Systems and Computers, 2009..

[21] B. F. Cron and C. H. Sherman, “Spatial-correlation functions for various noise models,” Journal of the Acoustical Society of America, vol. 34,38,

no. 11, pp. 1732–1736,885, November 1962.

[22] F. C. Robey, “A covariance modeling approach to adaptive beamforming and detection,” Ph.D. dissertation, Washington University, St. Louis, MO, Dec. 1990, also published as MIT Lincoln Laboratory Technical Report 918, 30 July 1991. Available through DTIC. Accessed: 2025-03-

08. [Online]. Available: https://apps.dtic.mil/sti/tr/pdf/ADA241887.pdf

[23] J. S. Chen, "Relative amplitude and phase features for resonance-region radar target identification," Proceedings of the IEEE National Aerospace and Electronics Conference, Dayton, OH, USA, 1989, pp. 97-101 vol.1, doi: 10.1109/NAECON.1989.40197.

[24] F. Xiang, S. Donglin, L. Yan and Z. Guoqi, "High-Frequency Radar Signature Simulation of Complex Targets," 2006 7th International Symposium on Antennas, Propagation & EM Theory, Guilin, China, 2006, pp. 1-4, doi: 10.1109/ISAPE.2006.353562.

[25] G. X. Wu and W. B. Deng, "HF Radar Target identification Based on Optimized Multi-frequency Features," 2006 CIE International Conference on Radar, Shanghai, China, 2006, pp. 1-4, doi:

10.1109/ICR.2006.343491.

[26] D.S. Ilcev, Introduction to coastal HF maritime surveillance radar. Polish Maritime Research, 26(Nr.3), pp. 153-162. 2019 DOI:

https://doi.org/10.2478/pomr-2019-0056

[27] J. Chen, Automatic Target Classification Using HF Multifrequency Radars. Ohio State Univ Columbus Electroscience Lab, Technical rept.

714190-3, October 1983,ADA162449

[28] S. J. Anderson, “Target Classification , Recognition and Identification with HF Radar.” (2004).

[29] S. Anderson, "Optimising bistatic HF radar configurations for target and environmental signature discrimination," 2007 Information, Decision and Control, Adelaide, SA, Australia, 2007, pp. 29-33, doi:

10.1109/IDC.2007.374520.

https://doi.org/10.1049/rsn2.12476 https://tc.copernicus.org/articles/14/1139/2020/ https://www.scribd.com/document/241494510/chaput https://doi.org/10.2478/pomr-2019-0056