Research Programs from BAA - Computing Sciences

1.1 Computational Architectures and Visualization

The Computational Architectures and Visualization program is concerned with modeling, analysis, and design of both hardware and software architectures with special emphasis on the effect the technological shift to heterogeneous, multi-core processors will have on these architectures. The program also supports innovative research in all aspects of visualization running the gamut from computational geometry and computer graphics to large-scale computer simulations and virtual environments. As the title suggests, the main thrusts of this program are computational architectures and visualization. Each thrust has specific areas of interest which can be found in the following paragraphs. However, an overarching theme for the program is the efficient managing and process sing of massive data sets. This is due to the fact that the Army's ability to generate data of all types from the battlefield to the laboratory far outpaces the Army's ability to efficiently manage, process, and visualize such massive amounts of information. The computational architecture thrust attempts to address this issue by sponsoring innovative architectural designs of both hardware and software components and their interfaces. The visualization thrust addresses the issue by providing support for innovative algorithms to render massive data sets and/or massive geometric models and to perform large scale simulations such as battlefield simulations, training simulations, and scientific simulations.

1.1.1 Computational Architectures

The goal of this thrust is to support the investigation of new effective architectures, computational methods and software tools for future computing systems. As computer hardware continues to improve and change, architectural modeling and design concepts (or paradigms) as well as software must adjust to take advantage of these improvements. These new computational systems must be scalable (usable on large-scale complex problems and able to handle massive amounts of data) and accurate (precise enough to predict and detect phenomena of interest). Other important issues to be considered for these advanced architectures, especially with the shift of computer manufacturers to multi-core processors, are programmability, language and compiler support, real-time scheduling, resource-allocation and the development of a flexible software environment.

1.1.2 Visualization

The visualization thrust of this program is concerned with all aspects of visualization and computer simulation of interest to the Army and is not limited to any one type of data or computer model. Specific research areas of interest are, but not limited to, discrete mathematics, computational geometry, robust geometric computing, graph theory, geometric and solid modeling, interactive graphics, 3D visualization tools, and synthetic environments. Special emphasis is placed on making very large simulations and the visualization of massive data sets faster, more computationally efficient, and more interactive for the user.

Technical Point of Contact: Dr. Joseph (Michael) Coyle, e-mail: joseph.michael.coyle@us.army.mil, telephone: (919)-549-4256

1.2 Information Processing & Fusion

Research is aimed at the development of theories, algorithms, and tools concerning data processing, information extraction, and information integration to support the development of real-time situational awareness and advanced targeting capabilities for military operations. Emphasis is placed on mathematical methodologies and algorithms for image processing, image understanding, video-based target recognition and tracking, and data/information fusion.

With the pervasive availability of unmanned systems in future military operations, advanced sensing will be of critical importance to the future force. This program emphasizes mathematical methodologies underlying automated sensing capabilities and robust target tracking through innovative approaches such as bio-inspired methods and optimized sensing asset allocation. Research efforts will support the creation of innovative algorithms for robust video-based tracking under challenging urban environments. Also supported is research on area monitoring using a network of cameras and other sensing modalities, potential applications including detection of improvised explosive devices (IEDs) and persistent surveillance. Research on brain-computer interfaces has the potential to provide a revolutionary capability for human-machine interactions. Work in this area actively supports advanced research in both minimally invasive and non-invasive imaging modalities for brain-computer interfaces.

The increased capability of electronic systems and proliferation of sensors are generating rapidly increasing quantities of data and information to the point that system operators and commanders are becoming saturated with information. Innovative methods for effective indexing and retrieval of multimodal data (e.g. imager and video data) are sought. An area of increasing importance is data and information fusion, especially from disparate sensors and contextual information. Sponsored research activities address several basic issues of data fusion, including information content characterization of sensor data and performance modeling. Of particular interest is research on collaborative target inference and target tracking by fusing data from disparate sensors as well as contextual information. Performance of target tracking can be improved through integrated approaches that consider target recognition, data association, and motion estimation.

Technical Point of Contact: Dr. Liyi Dai, e-mail: liyi.dai@us.army.mil, telephone: (919) 549-4350

1.3 Information and Software Assurance

From the Army perspective, Information Assurance must address the delivery of authentic, accurate, secure, reliable, timely information, regardless of threat conditions, over the distributed and heterogeneous computing and communication system. The computing system may range from a handheld mobile unit to a centralized high performance information process system. Heterogeneous communication system consists of both tactical (mobile, wireless) and fixed (wired) communications infrastructures.

1.3.1 Supporting Army Mission Characteristics

The objective of this research is to enable dynamic management of communities of interest (COI) by the battlefield commander. The commander needs the ability to alert the membership in a specific COI based on issues ranging from classification of the data to specifics of the battlefield situation. Individual war fighters may simultaneously be members of multiple COIs depending upon battle space specifics. Research is needed in the areas of protocols and techniques, which support reconfigurable, survivable and self-healing, efficient, and computing and communication environments that would allow for the dynamic creation of COI as well as to assure delivery of trustworthy data within reconfigurable and network centric environments. Reconfigurable, survivable, and self-healing systems allow a combat unit to dynamically establish and maintain its command and communication capability under diversified and extreme battlefield situations. New computing and communication protocols and techniques need to be developed so that critical information delivery and critical infrastructure functions can be assured, while maximizing the longevity of such systems under the resource constraints.

The Army requires a fully mobile, fully communicating, agile, and situationally aware force that operates in a highly dynamic, network-centric environment. This force consists of a heterogeneous mixture of individual Soldiers, ground vehicles, airborne platforms, unmanned aerial vehicles, robotics, and unattended sensor networks that operate in a complex wireless environment. To support net-centric warfare, new research is needed to 1) develop novel techniques for intrusion or anomaly detection and vulnerability assessment of mobile wireless networks that is automated, efficient, scalable, adaptive, and secure; and 2) develop security services and wireless security infrastructures for highly mobile tactical and unattended sensor networks that are distributed, scalable, and extremely resource efficient.

1.3.2 Innovative Approaches to Attack Prevention, Detection, and Response

Prevention works best when it is designed into the system architecture instead of being added on later, and so the development of system-security architectures and protocols that mitigate or reduce vulnerabilities is of critical importance. Research is needed to address the development and implementation of the security policies that these systems will need to support, as well as automated system configuration updates to allow a system to be flexible enough to meet a change in security policy or threat situation.

Formal representations coupled with cognitive learning-based approaches for zero-second identification and characterization are of utmost importance, potentially providing means of responding to unknown attacks based on automatically generated attack response processes, as opposed to employing the current signature-based state of the art. Research is needed on cognitive-based analytical techniques of attack response planning, wherein automatic processes are able to conjecture and determine response solutions to attacks, while providing confidence levels in the likelihood of correct attack categorization, and self-assessment of response impact to a host system's capabilities. Research on detection and mitigation of sophisticated cyber threats, enhancement of the security interoperability of new security technologies, and determination of hostile intent are sought in the following areas: 1) reasoning mechanisms supporting the identification, representation, and understanding of system, network, and application vulnerabilities by which security objectives are compromised, their origin, properties, and manifestation in software and hardware, and remediation; 2) development of techniques for detecting and depicting vulnerabilities using models, taxonomies, patterns, and representational tools (graphs, trees, etc.), including their structure and interrelationship of active and passive components; 3) diagnosis and analysis of attack mechanisms by which threats target our systems, networks, and information infrastructure, including study of preconditions and dependencies; 4) investigation and development of architectural strategies and solutions to counter potential security threats, using advanced methodologies and novel technologies; and 5) exploration, identification, and validation of metrics, measurement techniques, and probabilistic techniques by which the effectiveness of specific security solutions and compositions of security solutions may be characterized and differentiated.

Intrusion protection includes both host-based defenses that harden a host against attack and network-based defenses that identify and respond to problems identified in the network itself. New technologies will need to be developed that take into account the needs and special properties of emerging types of platforms such as wireless mobile devices. Innovative approaches to system protection are sought which will incorporate the development of anomaly based detection, correlation, and fusion methods, adaptive response mechanisms, and automatic generation of responses.

1.3.3 Development of Next Generation Resilient Computing and Communication Systems

A primary goal of this Information Assurance thrust is to define, develop, and evaluate systems and network architecture structures that would survive sophisticated attacks and intrusions. Another important goal of this research thrust is to define, develop, and evaluate systems and network architectures that scale to large configurations and yet retain the ability to self-heal and recover from unexpected events with measurable confidence. The objectives of this thrust are to 1) gain advanced understanding and knowledge of survivability principles for complex system design and development; 2) develop next generation communication and computing systems that are resilient against attacks of different levels of severity and are capable of recovering and self-healing from any potential compromises; and 3) develop benchmark and testing metrics to evaluate system integrity and survivability.

1.3.4 Embedded System Security

Embedded systems are used heavily in critical defense applications Malicious or accidental failures in embedded systems can have dire consequences. The integrity of embedded infrastructures, such as configuration and code, is of utmost importance. Another distinguishing feature of embedded systems is autonomous operation, which poses new challenges in the context of system integrity. Since embedded systems are reactive (interact with their environment), unexpected environment events can cause failures in embedded systems. Environment events can also be generated by a malicious adversary, whose goal is to have the embedded system fail. Novel techniques are needed to verify the integrity of embedded infrastructure (which may include device, software code, and configuration). These techniques should be geared towards discovering behavior of embedded systems under environment events and discovering "weak spots" or vulnerabilities in embedded infrastructures. Advanced techniques based on static, dynamic, or hybrid (a combination of static and dynamic) analysis of an embedded system are sought to identify exploitable vulnerabilities in embedded infrastructure (an exploitable vulnerability can be used by an adversary to drive the embedded system to an unsafe state) and to verify the integrity of embedded infrastructure.

Embedded systems often have to operate autonomously in a changing environment. Therefore, infrastructure of an embedded system has to be updated to adapt its behavior to the change in environment or change in the overall mission. Infrastructure updates manifest themselves as software updates or changes in the deployment configuration. Unauthorized or unverified updates to the infrastructure of an embedded system can compromise its integrity. New techniques are needed that allow updates to the infrastructure of an embedded system without violating its integrity. In addition techniques that prevent tampering with embedded infrastructures are also needed since field deployed embedded devices such as unmanned sensor nodes are prone to capture.

1.3.5 High Confidence, Robust, and Resilient Software

Today's software has unique characteristics of being large, complex, and decentralized. In addition, the current generation of software often operates in a heterogeneous environment, involving both infrastructure servers and mobile devices. To achieve the goal of information superiority in network-centric operations for the DoD, software for tactical systems must be highly reliable, adaptable, and flexible to changing environments, as well as resilient against potential attacks and intrusions. Research is needed in the area of high confidence, robust, and resilient software design, development, and verification, especially for mobile code and software supporting tactical mobile systems. The next generation software must provide full fault tolerance; can defend effectively against exploits and attacks; and be capable of healing and recovering from faults and potential compromises in order to sustain mission critical services. To achieve that goal, reliable and effective mechanisms to monitor and verify software execution status are also sought.

Technical Point of Contact: Dr. Cliff Wang, email: cliff.wang@us.army.mil, (919) 549-4207

 

Last Update / Reviewed: July 9, 2012