Artificial Intelligence

GCAS specializes in Artificial Intelligence techniques such as Expert Systems, Bayesian Networks, and Autonomous Agents, particularly for problems involving reasoning under uncertainty. Some recent projects include:

  • Advanced Planning and Scheduling Methods
  • Predicting the dynamic response and failure of electronic equipment subjected to shock and vibration
  • Expert Design Tools that Predict the deterioration of vehicles and systems due to Galvanic Corrosion
  • Advanced Search Methods for Digital Libraries
  • Development of the Physiological Models for the next generation military uniforms containing medical sensors
  • Extending Bayesian techniques to include second-order uncertainty
  • Development of Intelligent systems for planning Air/Battle Space under a Course of Action
  • Advanced Knowledge Acquisition Systems
  • Methods for predicting and tracking reliability/remain/maintenance of vehicles and systems
  • Data recording software for Aircraft, Ground Vehicle Maintenance
  • Software for recording traffic accident information
  • Prediction of Aircraft interior noise from Prop-fan engines
  • Development of Expert Design Tools for selecting paints and coatings for structures

Artificial Intelligence

Since our inception in 2000, we have won numerous Government contracts, including:

Corrosion Expert System (CES):

CES is a probabilistic expert system tied to embedded autonomous software agents in vehicle design and analysis software (CAD/CAE). The contract involved the development of a new expert tool for the early assessment of corrosion problems in a vehicle.

Interval Propagation in Bayesian Networks:

This MDA/Navy contract is to develop new algorithms for the propagation of second-order uncertainty (interval data) in Bayesian Network Models as stemming from imprecise knowledge (intervals in the numbers specifying the Conditional Probability Tables) and imprecise data (intervals in the evidence provided to the model from outside sources, such as standard deviation on a sensor measurement)

Warfighter Physiological Status Monitor (WPSM) Algorithms:

This current 2-year project involves the development of a set of Decision Support Algorithms for determining the alive/dead status of a soldier, methods for performing Probabilistic Risk Assessment (PRA) analysis of a suite of sensors to be worn by the soldier, a set of sensor models used to detect sensor failure and condition the measurements according to the sensor's reliability, and the development of a prototype Human Physiology and Sensor Simulator for modeling the combined human subject-sensors-WPSM controller system.

Search Methods for Scientific Data in Digital Libraries:

This NASA Ames contract involves the development of a new method of knowledge base information retrieval for large scientific databases, in which a request for information is posed as a question, and information sources are identified that pertain to steps in the logical process of answering the question.

Knowledge Acquisition for Level II/III Fusion:

This DARPA research project involved the design and development of new software tools for eliciting knowledge from Subject-Matter Experts (SMEs) having little or no training in knowledge representation and probability theory. The work concentrated on algorithms and methodologies for the acquisition and integration of mixed logical and probabilistic knowledge.

Blast and Shock Analysis for Vulnerability Codes:

In this Army/OSD contract, we are developing new prediction techniques using artificial intelligence methods for characterizing the failure of electronic equipment subjected to mechanical shock from blast waves created by conventional weapons.

Logic Model Development:

Under the National Institute of Health contract, GCAS is developing a computer software system that guides the development of Program Logic Models for program planning and evaluation. The software allows the user to develop a project plan, including resources needed for program operation (Resources/Inputs) and strategies or processes involved in implementation (Activities) linked to outcomes, including process outcomes (Outputs); desired changes to the target audience as a result of the program (Outcomes); and long-term program goals (Impact). The advanced software features include Scheduling (Milestone Charts, Gantt Charts, Pert Charts, Critical Path Analysis, Critical Chain Method), Project Risk Management (Belief Networks, Influence Diagrams), Timeline Budgeting and Earned Value Management.

Battlespace Management Planning and Scheduling:

In this Air Force contract, we are designing an interactive and continuous planning system that can handle multi-level objectives together with their deadlines and priorities. This planning system will be able to adapt to real-time information updates (for example, changes in the status of resources, additional or revised mission objectives, revised priorities of objectives, or target lists). In addition, the system will assist the users in specifying their assumptions, situation analysis, and external factors. Finally, the system will be designed to support centralized planning and decentralized execution.

Intelligent Coatings Selection System:

In this Navy contract, GCAS is developing a methodology for the retrieval of corrosion data from within existing databases and the framework for an Expert System that will model system breakdown utilizing existing coating data (in-service and laboratory testing).