A Portable Tactical Field Sensor Array for an Infrasound Direction-Finding and Positioning System

Full text

A Portable Tactical Field Sensor Array for an
Infrasound Direction-Finding and Positioning
System
John P. McIntire, U.S. Air Force Research Laboratory, 711th Human Performance Wing
Duy K. Nguyen, U.S. Air Force Research Laboratory, Aerospace Systems Directorate
Eric T. Vinande, U.S. Air Force Research Laboratory, Sensors Directorate
Frederick C. Webber, U.S. Air Force Research Laboratory, 711th Human Performance Wing
ABSTRACT
Infrasound refers to sound frequencies below the threshold of human hearing, around 20 Hz or less. There are a variety of
natural sources of infrasonic emissions, including thunderstorms, avalanches, meteors, earthquakes, volcanos, windstorms, etc.;
as well as man-made sources of emissions, such as aircraft, heavy machinery, artillery, missile testing, road traffic, etc.
Infrasound is especially attractive from a sensing perspective due to its ability to propagate long distances while suffering little
from atmospheric or environmental attenuation. In this work, we describe the development of a man-portable “tactical”
infrasound field sensor array that is small, lightweight, and can be rapidly set-up and torn-down as needed. The system is able
to provide direction-finding capabilities to infrasound impulse sources with a directional accuracy of +/- 3 degrees. Such
information could be used for alternative positioning schemes, as will be described in detail, or perhaps for direction-finding
(homing) to acoustic sources of interest. Possible users could be military or search-and-rescue teams operating in GPS-denied
environments; field researchers studying volcanology or seismology; or other geo-acoustic scientists and engineers.
INTRODUCTION
Infrasound refers to the “sub-sonic” region of the acoustic spectrum, consisting of sound frequencies below human hearing,
which by convention is defined as frequencies of 20 Hz or less. Due to its low frequency (and thus, long wavelength), infrasound
notably suffers very little atmospheric or environmental attenuation and so can propagate much longer distances relative to
higher frequency sounds. Distance of propagation depends for the most part on intensity of the sound pressure waves. Indeed,
powerful explosions can be detected almost anywhere on Earth. For instance, the Krakatoa volcanic eruption in 1883 and the
Great Siberian Meteorite of 1909 were detected all across the world on sensitive barometers, sometimes showing evidence of
the pressure waves circling the globe multiple times (Bedard & Georges, 2000).
All sounds suffer from spherical spreading effects as well as atmospheric and environmental attenuation, but infrasound suffers
considerably less from these latter effects, and the lower the frequency, the more apparent this propagation advantage becomes
evident. For instance, a 1000 Hz tone loses 90% of its energy due to atmospheric absorption after traveling 7 km, while a lower-
frequency 1 Hz wave can travel 3000 km before suffering equivalent degradation (Bedard & Georges, 2000). Above “classical
infrasound” frequencies, sounds up to 100 Hz can still travel considerable distances, particularly if the atmospheric and wind
conditions are favorable (Stubbs et al., 2005). Given that sound pressure levels (SPL) useful for infrasound sensing are typically
around 75 dB or higher (Stubbs et al., 2005), and sensor noise floors are commonly around 65 dB, an acoustic wave with
intensity of 160 dB SPL can carry for 30 km before dropping into the 70 dB range. The propagation distances of infrasound
can be truly astounding, particularly for the lower frequency, higher intensity sources.
Natural sources of infrasound emissions include sea waves, avalanches, wind turbulence, tornados, thunder, volcanos, meteors,
earthquakes, microbaroms (ocean wave noise), auroral activity and magnetic disturbances at polar regions. Some animals, such
as whales and elephants, use infrasound for communication and possibly navigation (Atlmann, 2001; Bedard & Georges, 2000).

Artificial or man-made sources of infrasound include aircraft engines, aircraft wake vortex and turbulence, helicopters, artillery,
blasting, heavy machinery (compressors, crushers, furnaces, etc.), heavy vehicles, ship engines, road traffic, rocket launches,
wind turbines, nuclear missile explosions, bombs (Altmann, 2001; Bedard & Georges, 2000), and perhaps underground
factories or facilities. Due to the wide variety of natural and artificial sources of infrasound, methods of detecting and studying
infrasound are of interest for both military and civilian applications.
Historically, infrasound sensing systems have been used extensively for the global monitoring of international compliance with
weapon test ban treaties, especially the United Nation’s CTBT: Comprehensive Nuclear-Test-Ban Treaty (Stubbs et al, 2005).
Low frequency sound has also had military interest as possible non-lethal acoustic weapons and active denial / crowd control
systems (Altmann, 2001) and long-range hailing and communication devices. Military use of infrasound has also focused on
long-range detection and direction-finding to air or ground vehicles that produce distinct (low-frequency) acoustic patterns,
like helicopters (Stubbs et al., 2005), tanks or trucks (Kaushik, Nance, & Ahuja, 2005), and since World War I, to detect and
locate enemy artillery fire (Altmann, 2001). Outside of a military context, infrasound arrays are also commonly used to study
and monitor volcanic and seismic activity and weather patterns including hurricanes, tornadoes, and atmospheric turbulence
(Shams et al., 2008).
For the most part, infrasound sensor systems are fixed and infrastructure-heavy permanent arrays, often of considerable size
(several kilometers between sensor elements); e.g., the international monitoring stations of the CTBT. Even the US Army’s
infrasound systems used to detect explosions, vehicles, missile launches, and underground facilities consist of sensor clusters
spaced 30 km apart and seem to be permanent or semi-permanent fixtures at specific locations, with sensing ranges up to 100
km. Tactically-deployable infrasound sensor systems were notably absent as of Stubbs et al. (2005) review, although low-
frequency (30 to 375 Hz) mobile/tactical acoustic field systems apparently have been used for helicopter detection and tracking
with ranges up to 20 km. The actual mobility or tactical portability of this latter system was not made explicit.
More modern efforts at portability are reflected in the work by Qamar Shams and colleagues at NASA Langley Research Center
(Shams, Zuckerwar, & Sealey, 2005; Shams et al., 2008) to develop a portable infrasound system requiring a vehicle for
transport but allowing for comparatively quick setup in the field; and by the small infrasound sensors designed to be lightweight
and man-portable offered by Chaparral Physics (in particular, the Model 60 series). Some additional engineering effort seems
necessary to adapt the current state-of-the-art in “portable” infrasound systems into a truly man-portable tactical system that
could be used in military field settings by small teams performing rapid-setup and tear-down of equipment that is small,
lightweight, wireless, low cost, with low power and computational requirements, and small geographic footprint. We attempted
to design and test such a custom system while utilizing commercially-available sensors, hardware, and software when possible.
A MAN-PORTABLE INFRASOUND FIELD SENSOR SYSTEM
The remainder of this paper will describe our man-portable infrasound sensor field array meant to be used within a tactical
environment by small military teams, to accomplish direction-finding and positioning for navigational purposes. The system
was developed as part of a larger set of alternative navigation tech solutions within a year-long international collaborative
research and innovation effort. Details of the project and some of its other technological outputs can be found in McIntire et al.
(pending) and Webber et al. (2016).
Infrasound Sensors. The sensors used in our prototype system were simple, lightweight (2 lbs), low-cost, commercially-
available differential air pressure sensors: a microbarograph design with solid-state differential pressure sensors and high-pass
pneumatic filter (Infiltec INFRA-20 Infrasound Monitor). The sensors were designed specifically to sense infrasound at 25 Hz
or below. The sampling rate is specified at approximately 50 Hz, and resolution is 0.001 Pascals over the range of +/- 20
Pascals. The hardware outputs to a serial cable, which was attached to a PC through a USB-serial adapter. Although we utilized
differential pressure sensors for our infrasound detectors, the use of infrasound microphones may also be possible as alternative
sensing devices.

Each sensor was outfitted with a several-meter-long hose to accomplish basic physical wind filtering. To provide additional
wind protection, we attached a foam “tip” to the end of the sensor hose. Each sensor is housed in small portable zip-up bags
with carrying handles that provide for thermal, wind, moisture, and physical protection of the main sensor unit box. See Figure
1.
Figure 1. Left: An infrasound sensor (Infiltec’s INFRA-20 sensor), with main sensor box, wind filter hose, and serial-to-USB
connector cabling. A standard pen is shown for size reference. Right: Each sensor, hose, and cabling is housed in a small
carrying bag that provided for portability, as well as permitting wind, thermal, and physical protection of the main sensor.
Field Sensor Array Configuration. It should be noted that many variations on the sensor array geometry are possible, and the
configuration may utilize three or more sensors for direction-finding. For our prototype development, we chose an array in the
shape of a right isosceles triangle (with angles 90, 45, and 45), with the side spanning from Sensor 1 to Sensor 2 bearing due
magnetic North; a standard field compass is sufficient for confirming this alignment in the field. Sensor spacing between the
sides should also be physically measured (in our example, the two side lengths are 35 meters each). See Figure 2.
Manual setup of the array involves placing each sensor at its desired location, unfurling the hose from the carrying bag, and
running the USB cabling from each sensor to the computer. Once complete, the system can be powered on and prepared for
data collection and analysis.

Figure 2. The tactical field sensor array setup. The current system requires an isosceles right triangle shape. The direction
from Sensor 1 to Sensor 2 should be due North (0 degrees) and the direction from Sensor 1 to Sensor 3 should be due East
(90 degrees). The system will then output the estimated source direction as a real-world compass bearing.
Synchronizing and Pre-Processing the Data. Data collection involves running a Python script that opens all the sensor’s
serial ports for a specified period of time, and simultaneously collects the raw data text stream coming from the sensors,
regularly re-synchronizes them to correct for any sensor time drift, and finally outputs each collected stream to a separate text
file for analysis, before closing the port and ending the data collection period.
Next, an R program is used to find the peak signal amplitudes in each sensor sample. Inputs into the program include the sensor
data text files, ambient air temperature, and physical measurements of the array (e.g., triangle spatial configuration and side
lengths). The search parameters for signal peaks is modifiable on-demand (to ensure the peak found in each sample is the true
signal we are searching for). Signal filtering (low-pass, high-pass, or bandpass) is also available if sensor or wind noise has
degraded the samples. Once the appropriate peaks are found by the system, time-differences-of-arrival (TDOA) can be
computed, and real-world directional estimates corresponding to these time differences can be extracted from this process, as
described next.
Directionalization Computations. TDOA direction-finding systems use the physics principle that velocity multiplied by time
equals distance, and directionalizing from TDOA systems requires at least three sensors. The use of additional sensors and
arbitrary or complex array geometry may give more accuracy but with the increasing cost of computational complexity (Ahmed,
Wei, Memon, Du, & Xie, 2013) and perhaps, impracticality for a field-deployable system. Once each signal ‘peak’ location in
time is identified from each sensor stream, three TDOAs are computed from the three sensor pair combinations (may be positive
or negative, depending upon which sensor peaks occur first):
TDOA1,2 = Time of Impulse Arrival at Sensor 1 – Time of Impulse Arrival at Sensor 2
TDOA1,3 = Time of Impulse Arrival at Sensor 1 – Time of Impulse Arrival at Sensor 3
TDOA2,3 = Time of Impulse Arrival at Sensor 2 – Time of Impulse Arrival at Sensor 3

From these three TDOAs, and with the use of the estimated speed of sound, c, we can compute three distances (D):
D1,2 = c * TDOA1,2
D1,3 = c * TDOA1,3
D2,3 = c * TDOA2,3
These three distances are then used to compute six Angle-of-Arrival (AOA) estimates for each sensor baseline direction (B),
two for each of the three sensor pairs. Note that this formula was adapted from well-known TDOA basic formulas, and is easily
derived using elementary geometry and trigonometry:
AOA1,2 = +/- arccos ( D1,2 / B1,2 )
AOA1,3 = +/- arccos ( D1,3 / B1,3 )
AOA2,3 = +/- arccos ( D2,3 / B2,3 )
These AOA estimates can be added/subtracted from the corresponding compass angles measured in the real world (see section
“Field Sensor Array Configuration” above) to arrive at estimated compass bearings for the true impulse direction. Note that six
possible angles are given as possible results. This is due to inherent ambiguity of the direction of arrival: any given TDOA for
a sensor pair computationally generates two possible different source directions (except in the rare case of a zero degree AOA).
Obviously, assuming a single source, only one of those directions in the pair will be the ‘true’ source direction while the other
will be incorrect, with no mathematical way to definitively differentiate between the two. But with three pairs of sensors in the
real-world arranged with different baseline directions, the six estimates should show a cluster of three ‘correct’ estimates
indicating the true single source direction, while the other three estimates should be scattered separately from the cluster. Visual
inspection of the estimates can be utilized by the user to discard the three incorrect estimates, or mathematical cluster analysis
techniques available in the software can be utilized. Subsequent averaging of the three ‘correct’ estimates then results, finally,
in one single estimate of the impulse source direction.
Figure 3. An information processing flow diagram of the infrasound direction-finding system. See text for further details.

A schematic flow diagram of the information processing steps of the entire sensor and processing system is shown in Figure 3.
Alternative, more mathematically-advanced methods for computing source estimate directions from TDOAs are available in
the acoustic localization literature (e.g., cross-correlations, independent component analysis, approximate maximum likelihood
estimation, etc.). Estimates can be improved by precisely measuring sensor positions and by accounting for physical
environment characteristics that affect the propagation of infrasound (e.g., air temperature is a dominant factor in the speed of
sound), which are easily adjustable parameters in our current software.
Schemes for Localization. There are several different schemes one can utilize to estimate the current geospatial position of
the tactical array in the field, relative to a known source position. Alternatively, such a system could be used by a tactical field
team with known positions to directionalize to emitters at unknown locations, for homing guidance or positioning.
Figure 4. Two possible array localization schemes based on number and types of infrasound impulse source emitters
(“synchronized” in this context means that the emitter is time-synced in some manner with the array sensors).
Synchronization allows the time of source emission to be determined versus the time of detection, so that the speed of sound
could be used to estimate distance to the array from the known emitter position(s). In both cases, to estimate the current
position of the array, the position(s) of the emitter(s) must be known.
For positioning an array with only one emitter at a known location, both a direction and a distance must be estimated. Direction
can be estimated using the system as described above (utilizing TDOA between sensors). Distance could be estimated by
knowing the precise time that the impulse is emitted at the source and detected at the sensors (or if the source and the sensor
array were time-synchronized and able to communicate). See Figure 4 (left). For positioning with two or more emitters at
known locations, time synchronization between emitters and the array is no longer necessary (but may still help). The array
simply detects the directions of the impulse sources, and computes a position via resectioning (i.e., angulation). See Figure 4
(right).
Tactical CONOPS. Which direction-finding or positioning scheme to be utilized in the field depends on the situation: what
position information is already known, and whether coordination, synchronization, or other information exchange is possible
between emitters and the sensor array. Two tactical field concepts-of-operation (CONOPS) are described next, illustrating the
use of both sensor array localization schemes as described above:

 One Friendly Source; Time-synchronization is Possible between Source and Sensors. Special Forces Team Alpha is
a few miles behind enemy lines. There are no clear landmarks in sight but an external verification on their position is
needed; GPS is not working. Team Alpha sets up the infrasound field array, and radios to Team Bravo who is a few
miles away at a known location. Over the radio, they are able to time synchronize their watches, and coordinate that
Team Bravo will denote a large explosive at a precise time. When this explosion is detected by the array at Team
Alpha’s position, the system is able to use the time difference to compute the distance the sound traveled, as well as
the direction of the emission. Team Alpha now knows their position with confidence; they quickly pack up the gear
and move forward with their mission.
 Two Friendly Sources; Time-Synchronization is Not Needed between Sources and Sensors. Special Forces Team
Alpha is a few miles behind enemy lines. There are no clear landmarks in sight but an external verification on their
position is needed; GPS is not working. They quickly setup the infrasound array in the field, and radio to a friendly
artillery group that has large guns positioned several miles away at several fixed locations. The group fires a shot from
gun one, then from gun two. Team Alpha senses the data, confirms their position, packs up the gear, and moves
forward with their mission.
EXPERIMENTAL RESULTS
A variety of outdoor field tests were conducted with our prototype system. Two important attributes were important to test, in
order to characterize the performance of our system: (1) the distance of detection of far infrasound sources; and (2) the
directionalization capability of our array. Tests were conducted in open natural settings surrounded by light forests, in the mid-
western United States (various sites in Ohio and Indiana) during Spring and Summer seasons, with temperatures typically
around 70-90 degrees Fahrenheit, generally high humidity, and occasional light wind.
Distance of Detection. These field tests were conducted at Camp Atterbury, Indiana, in cooperation with the US Army’s 107th
Field Artillery Unit. The purpose of the data collection was to determine the maximum distances at which our system could
detect large blasts -- emanating from model 119-A2 artillery guns using 105mm shells. Sound intensity of the weapons was
not measured on site, but various sources estimate artillery fire to create peak sound pressure levels as high as 180 dB SPL
(e.g., Chaillet, Hodge, Garinther, & Newcomb, 1964).
Starting at a near location (0.15 miles or 0.24 km), we setup a single sensor and ran data collections to confirm that blasts
coming from the artillery were consistently visible as large amplitude spikes on the software amplitude traces (see Figure 5,
left). If so, we would move farther away and test again. Repeating this process, we were able to confirm our infrasound sensors
were detecting the artillery fire at ranges up to 5.22 miles (8.4 km). A “zoomed-in” trace of our sensor data at the farthest
distance is show in Figure 5 (right). Further distances of detection are likely but were unconfirmed due to time and space
restrictions on the test firing range. It is also worth noting that at these larger distances, the acoustic blasts were only faintly
audible to the human ear, but appeared quite prominently in our data traces.
Direction Finding Tests: At multiple sites, in both Indiana and Ohio, several directionalization tests were performed. The
tactical sensor array was setup in open field areas, and the test impulses used were vehicle door and trunk slams which display
quite prominently in the sensor traces from distances of 50-100 meters away from the array. For several tests, we also utilized
a dual 18-inch subwoofer loudspeaker system with a high-wattage amplifier connected to a signal generator, which allowed for
the creation of arbitrary waveforms available on demand (typically using a 15 Hz tone played at max volume for 1-3 cycles at
a time using the pulse feature; this generated a clear acoustic spike in our sensors). The sensor array was always setup in the
field in the configuration described in Figure 2 and in the text: a North & East right triangle with sensor separations ranging
from about 25 to 35 meters (this varied across test sites, depending on terrain and ease of user setup). Once the spatial
configuration of the sensors was entered into the system, impulses could be generated for testing. The sensor operator would
begin by starting the data collection Python script; next, another operator would initiate the impulse generator; last, the sensor
operator would stop the data collection sample.

Figure 5. Left: Detection of artillery blasts at a near distance (0.15 miles or 0.24 km) using a single infrasound sensor, with
the sensor amplitude trace over time shown on Infiltec’s Amaseis software data and visualization package, and using some
basic bandpass filtering (5 to 25 Hz). The spikes are clearly visible as high amplitude impulses in the traces, confirming
sensor detection. Right: blasts detected at 5.22 miles (or 8.4 km) are still detectable but additional signal processing or wind
filtering techniques may make these impulsive signals more prominent above the noise.
Once a sample was collected, it was immediately loaded into the custom direction-analysis software, resulting in a directional
estimate of the source. This estimate could then be checked for accuracy by using a lensatic compass from the center of the
sensor array, to the source. The true source direction and then the estimate were recorded. Tweaking of software values could
be conducted at this time, between sample collections, allowing for adjustments to the digital filtering (low-pass, high-pass, or
bandpass, and at what cutoff frequencies) or for adjustment to the peak signal search parameters. Such tweaking was sometimes
necessary due to variations in wind noise levels at different sites and on different days/times. Tests were then often repeated a
number of times until the system was performing at its best. We often moved the acoustic source to different locations at the
same site to vary the direction. Through repeated testing under a variety of humidity, temperature, and wind conditions, and
variations in source distance and direction, and across different testing sites and at different times, we were able to consistently
achieve an estimated accuracy of at best +3.0 degrees from the true source directions.
DISCUSSION AND CONCLUSIONS
In our experimental tests, we were able to confirm that our tactical portable infrasound field array system was able to detect
loud impulsive infrasound sources at large distances (up to 5.2 miles or 8.4 km), and able to directionalize to impulses with
accuracies as good as +3 degrees. Although a direction-finding capability of 3 degrees is decent for a prototype system, this is
likely insufficient for real-world use for localization: Table 1 clarifies how an error of only 1 degree would translate into
positioning error as a function of distance from the source. From a source 10 km distant at a known location, one’s position
could be estimated with a precision of several hundred meters. At even more distant locations, the positional error estimate
grows to a several kilometer-sized region.
Table 1. Estimated Positioning Error per Distance from the Source(s), assuming +1 degree direction error
Distance
10 km
20 km
50 km
100 km
Position Error
+ 175 m
+ 349 m
+ 873 m
+ 1.75 km
One of the biggest problematic factors affecting infrasound sensing is wind noise. To overcome this issue, several techniques
could be used: signal processing, physical wind guards or filters, multiple redundant sensors for spatial averaging (Bedard &
Georges, 2000), and wind hose and pipe filtering (Walker & Hedlin, 2010; Noble, Alberts, Collier, Raspet, & Coleman, 2014).
Due to fact that forested environments provide a natural source of wind protection, forests are therefore considered ideal

environments for infrasound array system locations. In any case, mitigating the problematic issue of noise (specifically wind
noise) is an area ripe for future research and development work, in order to improve acoustic and infrasound sensing systems.
If artillery are unavailable or impractical for sensing purposes, finding artificial or natural sources for testing and development
can be a real challenge. Infrasound source generation often requires physically massive elements in order to move large amounts
of air. There are few practical options for infrasound source generators or speakers (although, see Domen, 2003). Loudspeakers
may need to be coupled with horns for reaching high-intensity for very distant propagation (Altmann, 2001). See also Bedard
& Georges (2000) for a loudspeaker coupled with a horn mounted in a pickup truck, which was able to reach 20 km using a
100 Hz generated signal; this group also apparently made a spherical Helmholtz resonator tuned to range of 10-50 Hz, which
was detectable at 30 km. Large organs, pipes, sirens, gongs, or drums may be possible for these purposes. See also Shams et
al. (2005, 2008) for details on their custom-designed infrasound generator, though note many of these would not be considered
good impulse generators. Rotary subwoofers may also be possible as an infrasound generator source, though it is unclear if
these speakers are capable of producing the necessary intensity for distant propagation.
Although our system was designed to detect and directionalize to impulsive infrasound sources (like a blast), a steady-state
constant signal that exists over time (like a motor or speaker) would have been better specifically for detection purposes -- as
even a very faint steady source can be detected quite easily among noise if data can be collected for lengthy periods. For
example, generating and subsequently detecting a faint but steady 15 Hz tone over some period of time is quite simple, while
detecting a wideband impulse that only exists in time for a fraction of a second is more practically challenging in the presence
of noise. This makes directionalization a bit more challenging, then, as TDOA of phase differences (instead of impulse arrival
time differences) must be determined. This did not seem possible with our sensors and signal processing system, despite our
best efforts at achieving within a limited time and budget. Future work on this issue is recommended to determine feasibility
for a portable tactical array system.
To advance the accuracy and utility of such a portable tactical system as described herein, we recommend the following:
 Engineer the sensors so they are wireless (to improve setup and tear-down time).
 Build dedicated software for processing data, computing solutions and providing a user interface.
 Expand the sensing frequency range above 25 Hz, perhaps into the 100 or 200 Hz range (see Stubbs et al., 2005; for
detailed thoughts on this issue).
 Replace or augment the array with low-frequency microphones as opposed (or in addition) to air pressure sensors.
 Using more sensors than three or four if this does not unduly hamper its practicality in the field.
 Investigate alternative (and more advanced) signal processing techniques for sensing and directionalizing (especially
regarding signal phase detection).
 Explore innovative options for infrasound generators and study natural sources of infrasound that may be usable as
fixed “beacons” or landmarks for positional reference; this is the concept of infrasound mapping or spatial
fingerprinting. See novel work along these lines conducted by Bauer, Nollet, & Biaz (2014).
 Investigate options for physical wind-filtering, like hoses, hose arrays, physical filter shields, etc.
Advancing the science in these key areas may provide a new and innovative way to perform direction-finding and localization
in the field via portable infrasound sensor arrays. Such systems may be useful for a variety of military, commercial, and
academic purposes.
ACKNOWLEDGMENTS
We would like to acknowledge the support and give our thanks to AFRL SUSTAIN Project advisors, mentors, and fellow team
members who assisted and supported us in this endeavor. This manuscript has been approved for public release by the US Air
Force Public Affairs office. DISTRIBUTION A: Approved for public release: distribution unlimited. Cleared 12 Oct 2016,
88ABW-2016-5075.

REFERENCES
Altmann, J. (2001). “Acoustic weapons – A prospective assessment.” Science & Global Security, 9, pp. 165-234.
Bauer, R., Nollet, S., & Biaz, S. (2014). “A novel approach to non-GPS navigation using infrasound.” Auburn University,
Technical paper CCSE14-03, 11 September 2014. Available:
http://www.eng.auburn.edu/files/acad_depts/csse/csse_technical_reports/csse14-03.pdf
Bedard, A. J., & Georges, T. M. (2000). “Atmospheric infrasound.” Acoustics Australia, 28(1), pp. 47-52.
Chaillet, R. F., Hodge, D. C., Garinther, G. R., & Newcomb, F. N. (1964). “High-intensity impulse noise: A major problem.”
U.S. Army technical note 4-64. Available: http://www.dtic.mil/dtic/tr/fulltext/u2/450726.pdf
Chaparral Physics. Model 60 series infrasound sensors. September, 2016. Specification sheet available at:
http://www.chaparralphysics.com/model60.html
Domen, J. K. (2003). “Infrasonic Helmholtz Resonator” United States Patent #6,665,413 B1, December 16, 2003.
Infiltec INFRA-20. The Inexpensive Infrasound Monitor Project. September, 2016. Software and sensor descriptions available
at: http://www.infiltec.com/Infrasound@home/
Kaushik, B., Nance, D., & Ahuja, K. K. (2005). “A review of the role of acoustic sensors in the modern battlefield.” Proceedings
of the 11th AIAA/CEAS Aeroacoustics Conference (26th AIAA Aeroacoustics Conference), 23-25 May 2005, Monterey,
California.
McIntire, J. P., Webber, F. C., Nguyen, D. K., Li, Y., Foong, S. H., Schafer, K., Chue, W. Y., Ang, K., & Vinande, E. T.
(pending). “LeapFrogging: A technique for accurate long-distance ground navigation and positioning without GPS.”
Accepted for presentation at ION (Institute of Navigation) Pacific PNT conference, and publication in the proceedings,
April 2017.
Noble, J. M., Alberts, W. C. K., Collier, S. L., Raspet, R., & Coleman, M. A. (2014). “Wind noise suppression for infrasound
sensors.” US Army Research Laboratory, Adelphi, MD. Tech Report ARL-TR-6873, March.
Shams, Q. A., Burkett, C. G., Zuckerwar, A. J., Lawrenson, C. C., & Masterman, M. (2008). “A portable infrasonic detection
system.” National Aeronautics and Space Administration (NASA) Langley Research Center, technical report. Accessed
Sept. 1, 2016. Available online at: http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/20080034649.pdf
Shams, Q. A., Zuckerwar, A. J., & Sealey, B. S. (2005). “Compact nonporous windscreen for infrasonic measurements.”
Journal of the Acoustical Society of America, 118(3), pp. 1335-1340.
Stubbs, C., Brenner, M., Bildsten, L., Dimotakis, P., Flatte, S., Goodman, J., Hearing, B., Max, C., Schwitters, R., & Tonry, J.
(2005). “Tactical Infrasound.” The MITRE Corporation, JASON group, Technical Report JSR-03-520. McLean, Virginia.
Walker, K. T., & Hedlin, M. A. H. (2010). “A review of wind-noise reduction methodologies.” Chapter 5, pp. 141-182 in
INFRASOUND MONITORING FOR ATMOSPHERIC STUDIES, A. Le Pichon, E. Blanc, & A. Hauchecorne (eds.).
Springer Netherlands.
Webber, F., Schafer, K., Vinande, E., McIntire, J. P., Jensen, D., Foong, S., Chue, W. Y., Li, Y., Nguyen, D., Ang Zong Yao,
K., Miller, M., Crawford, R., Wong, G. H., & Wood, K. (2016). “SUSTAIN international science and technology
collaboration: Methods & best Practices for non-co-located engineering project teams.” International Forum, American
Society for Engineering Education, 25 June 2016.