| Interactive Systems Modeling and Simulation Laboratory |  |
|
|
 |
|
Projects |
|
||||||||||||||||||||||||||||||||
|
As the Air Force (AF) moves towards becoming an Expeditionary Aerospace Force (EAF) composed of Air Expeditionary Forces (AEFs), logistical issues, for both deploying and sustaining these forces, will drive the EAF vision. Models of logistics processes and integrated information technologies to support distributed logistics personnel are critical in supporting the EAF concept. Current logistics models cannot address the myriad of foreseen and unforeseen issues that the EAF concept presents. The proposed research program will develop a web-based, portable, and reusable environment for simulation modeling and analysis to provide the timeliness and relevance required to conduct the necessary analyses facing Air Force logisticians, planners and strategists. The overall goal is to develop model-based information technology systems that support predictive analysis using web-based simulations and tailored user interfaces to logistics personnel at different levels of abstraction. The simulations and associated interfaces will support what-if analysis, look-ahead, and integration with real-time information from distributed sources. The proposed effort will: (1) develop theories and methodologies for integrating simulations with information support, (2) design, implement, and evaluate proof-of-concept analysis tools, (3) interface with ongoing efforts in the related programs in AFRL such as LOCIS that focuses on integrated technologies for information access and integration, and (4) support operational and staff logistics development and planning organizations. The overall approach will be supported by a computational architecture implemented using Java. Click here to enter the Air Force Logistics Modeling website
|
|
In the most basic sense, this research intends to examine the nature of human collaboration with the goal of gaining a better understanding of the cognitive processes involved. In the most practical sense, the research is to build systems that facilitate human collaboration and to model on computers the behavior that leads to successful cooperation. To study collaboration, we will study a particular type of collaboration that is relevant to many intelligent activities; that is, we will examine planning by teams of people that are engaged with computer systems and vast amounts of information. One of the most difficult tasks during such activity is to present to each person just the right amount of information in the form best suited to their particular piece of the team effort. To provide solutions to this problem, we will experiment with mixed-initiative approaches that synthesize human and machine decisions, case-based approaches that reuse past experience, agent-based technologies that encapsulate specialized knowledge, and cognitive systems engineering methods of user-modeling. |
|
This research sponsored by the US Air Force Research Laboratory is focused on the development of a computational architecture for modeling and simulation of future uninhabited aerial vehicles (UAVs). The architecture, called UMAST, being developed in this effort is useful to emulate characteristics of UAVs for studying human/system interaction. UMAST, implemented in Java, is extensible and facilitates concurrent, multi-user interactive simulations through a web browser with high-fidelity model-based user interfaces. Click here to access the UMAST website
|
|
Distributed modeling and simulation architectures that emulate complex, real-world dynamics and constraints are useful in the analysis and understanding of adaptive agent behavior, both human and automated, in the control of complex, dynamic, supervisory control systems. This research, sponsored by the US Air Force Research Lab, has resulted in the design of a distributed simulation environment and application in developing an adaptive agent architecture in the context of a multiple interconnected power plant facilities. The model provides an interconnected work domain that requires adaptive behavior from an operator. The architecture applies the model-view-controller paradigm in decomposing the simulation model from the user interfaces and systems control. The specific model is multi-layered in that it encompasses a business module, system control module, and a maintenance module. The modules are implemented as interdependent Java processes. The separation of these components allows the simulation to be distributed across both local and wide area networks that reflect the real-world complexity. In addition, the distributed simulation architecture provides for the modules to be controlled separately while the adaptive behavior of the operator is analyzed. The operator of the simulation may either be a human operator with supervisory control of the simulation through the primary system control interface or an intelligent agent process that is separated from the models. This intelligent agent embodies a generic architecture for adaptive human behavior and applies multiple knowledge representation methods.
|
|
This research supported by Systran Federal Corp., focused on the design and implementation of a distributed modeling and simulation architecture for representing printed circuit card systems. Using the architecture, an analyst can represent PCCA systems at different levels of abstraction including entities at the shop level, machine level, and at the machine sub-component level. The model layers can also be distributed where each modeling layer can be executed on different computer platforms. Real-time and dynamic data communication occurs during simulation execution where some of the objects can be synthetic objects and some updates could occur from real-world sensors. Thus, some of the data could be large and but not time critical, whereas other data could be small but time critical. A process planner can perform what-if analysis on different alternatives using a system interface at a remote computer and predict realistic system performance using several measures including cycle time and system throughput. The architecture implements concepts from object-oriented programming, model-view-controller paradigm, and distributed computing to provide a flexible environment to model several classes of placement/insertion machines and various PCCA systems. Parts of the simulation model are distributed across computing platforms. A placement controller encapsulates the control/coordination issues in model integration. A human process planner can connect in to the executing simulation and interact with the simulation through a graphical user interface. During simulation execution, data used by the simulation can change dynamically. There is significant inter-process communication between the distributed simulation models and between the simulation and the interface. The communication between computers is facilitated through Common Object Request Broker (CORBA) technology.
|
|
Information analysis involves discovering
information, determining contextual relevance, and synthesizing useful
knowledge from distributed, heterogeneous information sources. For
example, a logistician may potentially have to review a large quantity
of data from many relevant sources in order to make a resource decision.
During the review process, logisticians may be able to provide a list of
keywords related to the specific information that they need. The search
process then begins using this list of keywords. The logistician must
then manually filter the results of the search in order to synthesize
useful information. A search process based solely on keywords, may
result in missing important sources of information, finding too many
unrelated sources of data, or a combination of both situations due to
the problems of synonymy or polysemy. Synonymy is when several different
terms have the same meaning and polysemy is when a single word has more
than one meaning. Information about the content of the information
sources is needed for integrating sources and for overcoming the
synonymy and polysemy problems associated with key word search
techniques. Higher-level understanding of terms in a domain, i.e.,
semantics, can help in overcoming these problems. This higher-level
information can then be used for supporting information search and
retrieval techniques. Developing an understanding of information sources
in the form of semantic representation is fundamentally a complex
reasoning process. Manual generation of semantic concepts in a domain is
inefficient. Complete automation of the process could lead to
inaccuracies and therefore be ineffective.
|
|
Designing effective electronic information systems to support search and retrieval in real-world applications is a complex and highly challenging endeavor. Several factors, including the user's cognitive capabilities and limitations, the work domain and task constraints, and the content and form of the electronic medium, influence the interactive information-seeking process. This research, supported by LEXIS-NEXIS, developed an approach to model real-world information seeking in a corporate environment. Our approach used several methods, including self-reported user questionnaires, field studies of real-world professionals searching for information, and cognitive modeling techniques. The modeling results characterize different categories of corporate information seeking, goals of information seeking, electronic sources used, and search strategies and tactics applied during prototypical information search and retrieval. The modeling results were applied to the design of human-centered interfaces used in a real-world electronic information system.
|
|
Portable interactive simulations that support concurrent multi-user interaction are useful in the analysis of large and semi-structured problems especially those in which human interaction are of important consideration. Emerging concepts in distributed computing provide enabling technologies for the development of flexible and portable interactive simulation infrastructure. This research, supported by US Air Force Research Laboratory and Air Force Office of Scientific Research, led to the development of an object-based interactive simulation architecture, called JADIS and a web-enabled version JADIS-WEB, through which multiple users can interact concurrently with an executing simulation from distributed locations. The simulation executes on a server and users can connect in to the simulation and provide user commands through interfaces implemented as separate processes. The architectural components were applied to a real-world problem in the domain of airbase logistics. This research also evaluated the effectiveness of various means of facilitating communication between interfaces and the simulation process including the use of Java application programming through sockets, Java programs integrated with Common Object Request Broker Architecture technology, and Java's remote method invocation protocol. |
|
This research supported by Systran Federal Corp., focused on the design and implementation of a distributed modeling and simulation architecture for representing printed circuit card systems. Using the architecture, an analyst can represent PCCA systems at different levels of abstraction including entities at the shop level, machine level, and at the machine sub-component level. The model layers can also be distributed where each modeling layer can be executed on different computer platforms. Real-time and dynamic data communication occurs during simulation execution where some of the objects can be synthetic objects and some updates could occur from real-world sensors. Thus, some of the data could be large and but not time critical, whereas other data could be small but time critical. A process planner can perform what-if analysis on different alternatives using a system interface at a remote computer and predict realistic system performance using several measures including cycle time and system throughput. The architecture implements concepts from object-oriented programming, model-view-controller paradigm, and distributed computing to provide a flexible environment to model several classes of placement/insertion machines and various PCCA systems. Parts of the simulation model are distributed across computing platforms. A placement controller encapsulates the control/coordination issues in model integration. A human process planner can connect in to the executing simulation and interact with the simulation through a graphical user interface. During simulation execution, data used by the simulation can change dynamically. There is significant inter-process communication between the distributed simulation models and between the simulation and the interface. The communication between computers is facilitated through Common Object Request Broker (CORBA) technology.
|
|
In dynamic situations where task requirements and operator states can change from moment to moment, such as those found in battle conflicts or agile work environments, a fixed interface only provides the best mappings between the user and technology over the narrow predefined range fixed during design, producing sub optimal performance outside of this design envelope. The research aim of this project is to develop and test operator-system interfaces which adapt to the operator's state in real-time and thus provide the operator with the right information at the right time. Providing this capability will enable operators to make faster and more accurate decisions. The performance enhancement provided by adaptive interfaces will reduce the number of required operators, as well as reduce training requirements by enhancing the novice operator's performance ability. Previous research in adaptive interfaces has focused primarily on identifying and measuring operator state variables that could prompt adaptation of interface features and dynamic function allocation in noncomplex task environments. Click here to enter the Self-Evolving Adaptive Interfaces website |
|
Uninhabited Combat Aerial Vehicles (UCAVs) refer to UAVs capable of delivering weapons. The UCAV weapon system will exploit the design and operational freedoms of relocating the pilot outside the vehicle to enable a new paradigm in aircraft affordability while maintaining the rationale, judgment, and moral qualities of the human operator. The proposed research program develops a model-based approach to support human-centered automation in the command and control of remotely operated vehicles.. The models will be validated using expert planners and skilled pilots in the USAF. The computational models generated from this approach will define the information requirements and content for the human teleoperator and provide a testbed for what-if analysis. The proposed research adopts a multidisciplinary research approach bringing together experts on intelligent flight control systems, human interface design, and systems modeling in two DAGSI schools (WSU & AFIT) and across two branches (HECP & HECI) of AFRL's Human Effectiveness Directorate. Click here to enter the Command and Control of Remotely Operated Vehicles website
|