Acoustic Multi-Source Sensors Aid Geospatial Situational Awareness
- Details
- By Kevin Corbley | 10 December 2010 | 2885
In the arena of covert intelligence and surveillance, a key aspect of protecting critical infrastructure and securing vulnerable borders is to identify a potential threat before it knows it has been detected. This shifts the element of surprise in favor of the responders, who can neutralize the threat before it has a chance to inflict damage. Moreover, an accurate identification of the threat ensures the level of response will be adequate.
In the arena of covert intelligence and surveillance, a key aspect of protecting critical infrastructure and securing vulnerable borders is to identify a potential threat before it knows it has been detected. This shifts the element of surprise in favor of the responders, who can neutralize the threat before it has a chance to inflict damage. Moreover, an accurate identification of the threat ensures the level of response will be adequate.
The challenge of safe-guarding high-risk facilities, especially in remote areas, has become increasingly important to both U.S. government and private sector organizations since the 9/11 terror attacks. Aside from being enticing targets, military bases, maritime ports, oil refineries, and gas pipelines all have something else in common – miles of vulnerable perimeters that are difficult, if not impossible, to monitor with boots on the ground.
“Human beings are ideal in security positions because the human intellect is capable of instantly identifying and categorizing threats and selecting appropriate actions,” said Alex Philp, PhD, President and CEO of TerraEchos Inc. in Missoula, Montana. “The problem is that it’s impractical to deploy people every few feet along miles of political borders or facility perimeters.”
In 2007, Philp spun TerraEchos out of another high-tech company, GCS Research, specifically to develop a covert sensor fusion technology with the ability to detect and locate potential security breaches and apply intelligent algorithms to identify those threats with a high level of probability. The result is the Adelos S4™ system, which this year received awards from IBM and Frost & Sullivan for its innovative integration of acoustic fiber-optic sensor technology, multi-stream information processing, and geospatial situational awareness.
Inside Adelos
Adelos, which means ‘hidden’ in Greek, leverages an advanced fiber-optic acoustic sensor technology developed by the U.S. Navy at the Naval Undersea Warfare Center in Newport, R.I., for use in protecting shipping ports. The thin fiber-optic cable can be deployed in segments measuring miles in length, either submerged underwater or buried underground to serve as “ears” that detect noises. Depending on the environment, these noises could be the sounds of approaching ships, vehicles, animals or human beings.
 |
| This cross-section diagram illustrates the components of the Adelos fiber-optic acoustic sensor. |
“Detecting a noise has limited value until you can start to answer the ‘what,’ ‘where,’ ‘when,’ and ‘who’ questions,” said Philp. “And that’s where geospatial intelligence and advanced analytics become part of the solution.”
Using a miles-long gas pipeline as a hypothetical example, Philp explained that even before the fiber-optic cable is laid along the right of way, a GPS survey is conducted of the perimeter to be protected. In the case of the pipeline, the cable may be deployed along its entire length or in sections at particularly vulnerable areas. A digital GIS map is created for the deployment with the location of each acoustic sensor pinpointed with a high degree of spatial accuracy.
The sensors themselves are not actual pieces of hardware. They are virtual sensors created by transmitting laser light down the fiber to form zones that are sensitive to acoustic energy. The width and spacing of the acoustic zones can be customized for different surveillance situations.
When a sound – traveling through the air, ground or water – disturbs the fiber zone, the detected noise is noted by the Adelos analytics engine at the terminus of the cable, and the sound location is georeferenced to the sensor that detected it and mapped to the GIS. If the noise is deemed a possible threat, the end user knows precisely where along the pipeline to send a responder.
This answers ‘when’ and ‘where’ the noise was detected, but significant additional value lies is determining ‘what’ and even ‘who’ made the sound. In a pipeline corridor, however, there are thousands of sounds, and they usually don’t occur one at a time. The normal flow of gas or oil through the pipeline makes one sound, while wildlife in the background makes many others. And an intruder climbing onto the pipe might make a distinctly different noise.
“Parsing out the sounds that potentially come from legitimate threats is critical, and it’s what sets Adelos apart from other acoustic sensor networks,” said Philp.
Integrating Streams
The breakthrough in Adelos development came in 2010 with the integration of IBM InfoSphere Streams technology into the solution. As the first commercial licensee of Streams, TerraEchos placed the IBM technology as the processing engine under the hood of Adelos allowing the acoustic network to process 1600 MB per second of digital acoustic data streams as it flows in from the sensors.
“IBM InfoSphere Streams is a fusion engine built on scalable and parallel processing platform that can digest massive volumes of data in motion, and Streams is extensible to support intelligence gathering over multiple sensor networks simultaneously,” said Philp. “This enormous processing capacity gives Adelos two unique capabilities.”
The first of these is the ability to answer the ‘what’ question, explained Philp. Analytics algorithms in the system analyze the multitude of noises being fed in from the acoustic sensors in real time and differentiate them into individual sounds. Streams enables Adelos to run a sophisticated array of statistics in a parallelized way so they can be classified and identified. In the pipeline example, the sounds of normal internal gas flow, valve activities, outside wildlife calls and thousands of others would be excluded from a specific target list. While everything is detected given the incredible sensitivity of the fiber array, only threat targets are communicated to security officers for response action.
“Certain acoustic patterns would be labeled in the archive as anomalous or threatening while others would be attributed to normal operations,” said Philp. “For example, a sudden change in the flow of gas through the pipeline might indicate a mechanical problem, while a vehicle driving up to the pipe might signal an intruder.”
When a noise is classified as a possible threat or as unknown, the communications component of Adelos sends a brief alert to the operator, pinpointing the location of the noise and providing a description of what might have caused it, as well as providing a sound file of the target. Delivering such intelligence in near-real time may be sufficient for many covert surveillance applications, but for others it may not be smart enough. And for those uses, Adelos attempts to answer the ‘who’ question.
The second major advantage the Streams technology brings to Adelos is interoperability – it offers the ability to consume, digest and analyze data inputs from many different sources. Streams was not designed solely for acoustic data processing. It can handle both structured and unstructured data as diverse as images, video, email, voice traffic, GPS points, and a variety of biometrics. This makes it possible to take the acoustic information picked up by the Adelos sensors and cross reference it with intelligence from other sensors or other databases to move beyond merely detecting a threat to identifying it and possibly giving it a name.
In the pipeline monitoring example, the Adelos network would relatively easily detect and recognize the sound of human footsteps approaching the pipe. But the system can then send a trigger to a hidden surveillance camera to activate and swivel directly to the point on the ground where the footsteps were noted. At that point, the Adelos processing engine begins to analyze the live video feed from the camera and extract frames with clear views of the intruder’s face.
“In a matter of seconds, Streams can reach out to an appropriate biometric database and compare the facial features of the intruder against known persons of interest,” said Philp. “If a positive identification is made against a known threat list, that information will be conveyed to the pipeline operator, and it will help dictate the level and type of response needed to defend the infrastructure.”
Adelos has a built-in communications server to electronically issue alerts to end users. Delivered within seconds of detection by the sensors, these alerts contain details on the identification, location and assessment of the threat.
Adelos S4 has been deployed in several advanced research and development pilot projects by the U.S. government in the interest of perimeter protection and border security, said Philp. However, TerraEchos is also fielding inquiries from energy producers, utilities, pipeline operators and other private sector organizations with large facilities that require 24/7 surveillance.
About the Author – Kevin Corbley is a communications consultant specializing in the geospatial and location-based services profession. He may be reached at
This email address is being protected from spambots. You need JavaScript enabled to view it.
.
