Искуственный Интеллект:
Планы на 2010-2011 год: разработать алгоритхмы способные с нуля обрабатывать сенсорные данные, без исxодныx данныx о структуре окружающей среды
Поддержка боевой еффективности
Планы на 2010-2011: Раработка препаратов для убыстренной аклиматизации на высоте.
Когнитивная Систем Предупреждения Угрозы
Планы на 2010-2011: Разработать систему способную найти пеxоту и машины на радиусе от 1-10 км, в диапазоне 120 градусов. На данный момент продемоснтрированна спосбность найти цели в диапазоне 20 градусов, с вероятностию 98% и менее 10 ложныx тревог за менее 60 секунд оптического сканированя поля зрительности.В 2011-м году провести 6-месячные испытания в полевыx условияx.
Акустическая Коагуляция Глубокиx Ран
Автоматическая система наxодит, локализирует и коагулирует глубокие ранениея в боевой обстановке с минимально подготовленым оператором. Модуль размером в 2.5 кг разработан, в 2011 планируется уменьшить систему до 200 грам.
The Foundational Machine Intelligence program is supporting research on the foundations of
artificial intelligence and machine learning and reasoning. One focus is on techniques that can
efficiently process and “understand” massive data streams. Deeply layered machine learning engines
will be created that use a single set of methods in multiple layers (at least three internally) to generate
progressively more sophisticated representations of patterns, invariants, and correlations from data
inputs. These will have far-reaching military implications with potential applications such as anomaly
detection, object recognition, language understanding, information retrieval, pattern recognition, robotic
task learning and automatic metadata extraction from video streams, sensor data, and multi-media
objects. Foundational Machine Intelligence also examines the human aspects of computing, with
interest in collaboration, interaction and information exchange; non-symbolic representation/reasoning
paradigms based upon a universal “cortical” algorithm; unmanned vehicles and intelligent agents that
generate and manage their own goals within human-described mission constraints; and modeling of
human language acquisition by associating words with the real-world entities perceived through multiple
modes of sensory input.
FY 2010 Plans:
- Create machine learning techniques that can assimilate huge amounts of data by creating rich
representations of the input data and applying them to multiple applications.
- Construct a single, general-purpose algorithm which could start with zero knowledge of its
environment, and then grow to represent the structure latent in that environment.
FY 2011 Base Plans:
- Create parameter-free methods that learn appropriate representations starting from raw inputs with a
single architecture and learning algorithm.
- Enable machines to incorporate sensory information in a robust way to improve situational
awareness.
Maintaining Combat Performance
(U) The Maintaining Combat Performance thrust utilizes breakthroughs in biology and physiology to
sustain the peak physical and cognitive performance of warfighters operating in extreme conditions.
Today, warfighters must accomplish their missions despite extraordinary physiologic stress. Examples
of these stressors include extremes of temperature (-20 degrees F to 125 degrees F), oxygen
deficiency in mountains, personal loads in excess of 100 lbs, dehydration, psychological stress,
and even performance of life-sustaining maneuvers following combat injury. Not only must troops
maintain optimum physical performance, but also peak cognitive performance, which includes the entire
spectrum from personal navigation and target recognition, to complex command and control decisions,
and intelligence synthesis. The Maintaining Combat Performance thrust leverages breakthroughs in
diverse scientific fields in order to mitigate the effects of harsh combat environments. For example,
understanding the natural mechanisms for core body temperature regulation in hibernating mammals
has led to a novel, practical approach for soldier cooling, which is now being evaluated by troops in
the far forward combat areas. Other examples include fundamental research elucidating the biological
mechanisms of adaptation to extreme altitude, and the molecular correlates of muscle fatigue and
psychological stress.
Cognitive Technology Threat Warning System (CT2WS)
(U) Recent advances in computational and neural sciences indicate it is possible to push the visual
threat detection envelope to enable more response choices for our soldiers than ever before. The
objective of the Cognitive Technology Threat Warning System (CT2WS) program is to drive a
breakthrough in soldier-portable visual threat warning devices by leveraging discoveries in the disparate
technology areas of flat-field, wide-angle optics, large pixel-count digital imagers, visual processing
pathways, neurally based target detection signatures and ultra-low power analog-digital hybrid signal
processing electronics. This program will lead to the development of prototype soldier-portable digital imaging threat queuing systems capable of effective detection ranges of 1-10 km against dismounts
and vehicles. Simultaneously, the system will survey a 120-degree or greater field of view, enabling
the warfighter to detect, decide and act on the most advantageous timeline in complex operational
environments.
FY 2009 Accomplishments:
- Demonstrated single path (twenty degree by twenty degree) advanced optics on a breadboard
system in a field environment consistent with objective performance and package volume.
- Demonstrated human-in-the-loop integration with the breadboard system, harnessing non-invasive
neural signatures for threat detection.
- Demonstrated visual/cognitive algorithm performance for threat detection on operationally significant
image streams with probability of detection (greater than .98) and false alarm rates (less than ten) in
less than sixty seconds of scan time.
- Demonstrated composite software system capable of high fidelity threat detection with extremely low
false alarm rates.
- Tested breadboard performance during week-long operational test at Yuma Proving Ground, AZ.
FY 2010 Plans:
- Develop integrated brassboard designs consistent with desired threat cueing performance.
- Increase field of view to 120 degrees x twenty degrees while maintaining size, weight and power
constraints.
- Demonstrate visual/cognitive algorithm performance for threat detection on operationally significant
image streams with probability of detection (greater than .98) and false alarm rates (less than ten) in
less than thirty seconds of scan time.
- Complete critical design review of bench-integrated prototype system evaluations that demonstrate
the capability of the design to meet the objective system program performance.
- Evaluate device packaging approaches with the knowledge of ruggedization and robustness
required for soldier-portable tactical electronic devices.
Tactical Biomedical Technologies
(U) The Tactical Biomedical Technologies thrust will develop new approaches to deliver life-saving
medical care on the battlefield, as well as novel technologies for reconstruction and rehabilitation of severely injured warfighters. Implicit in this thrust is the fact that there are unique, warfighter-specific
challenges in acute and chronic treatment that are not addressed by civilian research and development.
Today, more than half of American battlefield fatalities are due to hemorrhage, particularly due to
improvised explosive devices (IEDs). To prevent these deaths, there is an urgent need for technologies
that enable relatively unskilled personnel (battlefield medics) to diagnose and treat injuries, including
the ability to locate and coagulate non-compressible deep bleeders in the thorax or abdomen. Other
critical needs stem from the fact that warfighters are frequently victims of blasts, causing patterns
of brain, burn, and orthopedic injuries not seen in civilian medical practice. As such, there is a
unique military need to develop systems for pain control that are safe even in medically unmonitored
environments, such as an active battlefield. Once lives are saved, there is an unmet need for new
methods to restore function, for example, by restoring long segments of bone that were lost due to blast
fragmentation. The results of this program will greatly enhance our ability to save lives on the battlefield
and provide restoration of normal function to survivors.
FY 2009 Accomplishments:
- Demonstrated extended survival time using an FDA-approved estrogen product after 60% total
blood volume loss in swine hemorrhage model.
- Developed a physiological-based pharmacokinetic/pharmacodynamic model of the cardiovascular
system to aid in determining appropriate estrogen doses in humans suffering lethal hemorrhage.
- Demonstrated blastemal associated initiation of early joint formation at appropriate site during
healing.
- The Deep Bleeder Acoustic Coagulation (DBAC) program is currently developing a portable, noninvasive,
automated system for the detection, localization, and coagulation of deep bleeders that is
operable in the combat environment by minimally trained personnel. The stationary wrap-around
device must prove to be lightweight and operate on batteries. To this end, one therapy module and
one detection and localization (D&L) module with weight commensurate to meet a full 40 x 80 cm cuff
weight of less than or equal to 4.8 kg was successfully designed and built.
- Conducted in vivo and in vitro experiments to determine the effect of physiological variables on the
DBAC algorithm.
- Developed and tested automated algorithms for bleeder detection, localization, coagulation, and cuff
control and integrated into a 2.4 kg prototype cuff.
- Identified two materials capable of infiltrating into both penetrating noncompressible wounds and
surface wounds for potential use in new wound-healing technology.
- Determined specific wound biomarkers for targeting hemostatic (stops bleeding) materials.
FY 2010 Plans:
- Demonstrate in vivo induction of restorative skeletal muscle repair by transplant of induced
pluripotent cells.
- Determine transition kinetics from joint formation to bone morphogenic protein-2 (BMP-2)-induced
long bone restoration.
- Develop a material that can be delivered to a closed, intracavity space and binds specifically to
damaged tissue as demonstrated in situ by immunohistology.
- Demonstrate that hemostatic material does not induce intracavity scar formation within 28 days
when left at the wound site.
- Build and demonstrate an automated laboratory prototype DBAC system.
- Optimize automated algorithms for bleeder detection, localization, coagulation, and cuff control with
in vivo models.
FY 2011 Base Plans:
- Demonstrate compatibility with FDA-approved agents that control pain, infection, and inflammation.
- Achieve wound treatment system unit specs including coverage of at least 0.20 square meters of
tissue area, mass of less than 200 grams, and a volume less than 150 ml.
- Demonstrate hemostasis in less than four minutes on a high-pressure non-compressible injury
model.
- Maintain hemostasis in high pressure model for three hours.
