Author Contact Information
Capt. Paul Harmer
AFRL/SNAS, Bldg. 23
Fifth Street
WPAFB, OH 45433-7001
www.vdl.afrl.af.mil
 |
 |
 |
WRIGHT PATTERSON AIR FORCE BASE, Ohio (AFRL/SN) — The mission of the Air Force Research Laboratory (AFRL) Sensors Directorate is to ensure unequaled reconnaissance, surveillance, precision engagement and electronic warfare capabilities for America’s air and space forces, by conceiving, demonstrating and transitioning advanced sensors and sensor technologies. In simple terms, we’re the "Eyes and Ears of the Warfighter". We accomplish this daunting task in partnership with industry, universities and other DoD agencies.
A recent announcement by the Department of Defense High-Performance Computing Modernization Office (HPCMO) provides the Sensors Directorate with one more tool to support our mission. The HPCMO announced that AFRL/SN is now an HPC Distributed Center and awarded the funds needed to purchase needed HPC resources. Acquiring HPC resources at AFRL/SN adds a much-needed capability to an ongoing program sponsored by the Office of Secretary of Defense (OSD) called the Virtual Distributed Laboratory (VDL).
The Virtual Distributed Laboratory (VDL) is an information distribution center that provides researchers the capability to talk about, share, store, search, and retrieve information related to Automatic Target Recognition (ATR) and Information Fusion technologies. The addition of HPC provides users with the computing power needed to process these technologies in real time. Currently, VDL supports 31 programs and over 800 users, including the Air Force, Army, Navy, DARPA and several other DoD level organizations. A more detailed explanation about VDL will be provided, but first we need to understand the goals of ATR and sensor fusion.
Remember, the goal of the Sensors Directorate is to be the "Eyes and Ears of the Warfighter". So, one question to ask is: "What does the warfighter (in this example, battlefield commander) need to see and hear?
If you have ever played the childhood game of Hide and Seek, you understand two of the basic needs of a battlefield commander. First, the commander has to find where the enemy is hiding. Next, he has to use that information to formulate a plan of attack. In hide and seek terms, he has to find where the people are hiding and tag them before they reach home base and are safe. In the game, the people you are trying to find have the advantage because the seeker has to hide his/her eyes and count to some number while the people hide. Similarly, in warfare, if the enemy can move around and hide while we can’t see him, he has the advantage. Our goal in the Sensors Directorate is to give the battlefield commander the ability to look while the enemy is trying to hide or to find him after he has hidden.
One way we do that is by developing and fielding remote sensors that can "see" and "watch" over an area of interest and provide that information to the commander in the form of images. Image analysts review these images to decide where the enemy is hiding and what he may be planning. A problem arises when you have too much image data and not enough image analysts to review it. Using the hide and seek example, it’s like being able to look while you count, but having to count to a million before seeking. In the game, the players who are hiding are likely to quit and go home before you finish counting to a million and begin to look for them. In war, if it takes too long to analyze and react to data, the information becomes useless. Therefore, the ability to see what the enemy is doing without the ability to analyze and to act on this information in a timely manner is useless. ATR technologies are designed to make timely analysis of massive amounts of sensor data possible.
Researchers design ATR technologies that automatically process remote sensor data using computer algorithms to extract, sort, condense, and fuse the relevant information in these images. The processed information is then presented to battlefield commanders, allowing them to make faster and better battlefield decisions. Therefore, the commanders not only get the ability to look while counting, but they also get the relevant information in a timely manner. This allows them to react while the information is still useful. The question is: How does VDL help to accomplish this?
VDL is a high-speed high bandwidth interactive network system that facilitates collaboration between nationally distributed DoD research and developers using the following five functional areas:
In addition, it allows honest-broker evaluation of ATR algorithms and data sets. In simple terms, VDL helps researchers and developers from all over the country work together to develop better ATR and Information Fusion technologies. With the addition of embedded HPC resources, it will also help to reduce the time it takes to get those technologies to the warfighter. Finally, combining VDL with a high-speed network foundation provides worldwide users with the very best in remote information management and high performance computing to directly support Information Superiority in the 21st Century.
One of the most critical steps to getting technology to the warfighter is making sure it can be fielded. Which means, it must fit in an aircraft platform, in a mobile ground unit, or on a ship. It also must support the shock, temperature, and power requirements that go along with those environments. Because of size, weight, shock, temperature, and power requirements, our HPC Distributed Center (DC) will use embedded HPC resources. Embedded High Performance Computers provide the CPU power needed to support the real-time requirements for Automatic Target Recognition and Information Fusion, while at the same time support the fielding requirements.
As stated, in future conflicts it will not be possible or adequate to throw more people at the problem. Battlefield awareness will require the acquisition, assimilation, and analysis of vast amounts of remote sensor data at rates that far exceed the capabilities of human analysts. ATR technologies can bridge that gap, directly supporting "Information Superiority". A distributed center dedicated to the real-time requirements for Automatic Target Recognition (ATR), signal/image processing (SIP), and Integrated Modeling and Test (IMT) can demonstrate greater than a 10x speedup in development and transition of critical ATR technologies to warfighters.
HPC resource centers allow technologies to be developed using similar, and in some cases, the same hardware currently deployed in the field. The VDL HPC center provides a capability that greatly reduces the cost of ATR research and development by lowering and eliminating duplication of efforts in data collection and retrieval, information distribution and data storage. It also provides a cost avoidance to organizations from engineers and scientists’ time spent conducting ATR, Fusion, and C4IRS literature searches that often result in no matching, overwhelming matches that necessitate detail reading of advertisements, and unrelated subject matter when using the internet super highway to information.
The VDL HPC center emphasizes that the center cannot store everything of interest to the DoD community. A huge amount of additional information is stored on other program, government, academic, and industry web sites. Commercial search engines are great for locating such distributed information. "Our problem has been that the public search engines index too much content for our purposes" stated by the Center’s team leader, Captain Terry Wilson. The VDL Central Library provides researchers the capability to quickly find pertinent information, and spend more time developing solutions and solving problems. This capability will result in providing faster technology transition to the war-fighters, and improve the effectiveness and efficiency of technical solutions to complex and recurring weapon system issues.
The addition of an interactive Real-Time High Performance Distributed Center to the VDL program provides the capability to develop and share high-performance interactive modeling and simulation tools and to create and evaluate advanced embedded ATR technologies. It also provides a development infrastructure that mirrors many of the embedded system hardware architectures currently in the field. Eliminating the need to port code or at least making it much easier facilitates rapid demonstration/transition of new technology to warfighters.
The bottom line is that the addition of a real-time classified and unclassified interactive HPC, for shared use by the DoD community, affords a unique opportunity to develop and share ATR technologies across the DoD and enables successful transition of HPC technology to the fielded production systems.
Capt. Paul Harmer
AFRL/SNAS, Bldg. 23
Fifth Street
WPAFB, OH 45433-7001
www.vdl.afrl.af.mil
|
|
|
|