| 2005 Technology
Symposium > Modeling, Simulation and Training
Modeling, Simulation and Training
This area focuses on information technology to support training, and
technology and innovative application of modeling and simulation. The
information revolution is fueling changes in the workplace at an unprecedented
rate, and these changes are threatening to overwhelm conventional education
and training approaches. Fortunately, advanced instructional technologies
like embedded training and collaborative learning environments can help
warfighters and intelligence analysts adapt to these changes. Advances
in simulation infrastructure, interoperability architectures, and modeling
paradigms, have simplified the application of simulation, demonstrated
the feasibility of building simulations from reusable components, and
otherwise facilitated a revolution in simulation application.
Airport Demand/Capacity Model
Ashley Williams, Principal Investigator
Location(s): Washington
Problem
The FAA is exploring market-based solutions to capacity and demand imbalances at several airports. Evaluation of these possibilities requires high-fidelity modeling of the likely responses of airport users. However, in an industry with multiple interdependencies and demand for products based on at least five choice dimensions, a closed-form solution is simply intractable.
Objectives
The project will develop the MarketFX model to facilitate policy-relevant predictions, such as changes to the average fares, number of carriers, and destinations served at the airport. MarketFX is a simulation of the envisioned airport environment that uses the latest machine learning techniques to model the profit-driven evolution of airport users' schedules.
Activities
This research will finalize the MarketFX model through a six-phase test plan, progressing from simple, easily validated test cases to complex, not easily validated ones. We will validate the core functionality under correspondingly simple test cases, then add and test more advanced functionality. When the test plan is complete we will analyze price-based operations at LaGuardia (LGA) and Chicago O'Hare (ORD).
Impact
We plan to deliver the following products as a result of this research: a finalized MarketFX model; technical analyses of proposed policies at LGA and ORD, targeted for delivery to the FAA; technical paper(s) describing unique elements of the model, targeted for publication in a peer-reviewed journal; and possible conference presentations of technical papers.
^TOP
Delay Influence Analysis
Jackie Kee, Principal Investigator
Location(s): Washington
Problem
In 2004 we began to see the return of high numbers of delays at several large airports in the National Airspace System (NAS). One challenge in analyzing airport delays is evaluating the interaction among these delays across multiple airports. The analysis of delay influence can provide valuable information to the FAA, especially when corrective or preventive action is required to improve on-time performance in the NAS.
Objectives
We will investigate the influence of delay across 35 major U.S. airports, specifically focusing on quantifying the impact of delays at one airport on the other airports. We will analyze historical Airline Service Quality Performance data to attribute the delays incurred on the ground and in the airborne environment to the appropriate airport and/or en-route environment. The analysis will also consider the influence of different weather days.
Activities
We will develop an algorithm to track and measure ground and airborne delay for 35 major airports. We will run it for a range of dates and analyze the results to characterize the delay influence at each airport. Scripts will be developed for analysts to use in investigating delay influence at any of the 35 major airports.
Impact
The algorithm will allow us to analyze delay influence systematically, via analyst-friendly scripts. As the aviation industry recovers from its four-year slump we expect to see the recurrence of delays at key points in the system. This analytic capability will position us to better understand the networked effects of these delays and to provide tools and analyses that help inform FAA decision-making processes in a timely manner.
^TOP
Java Meets Simulation
Richard Weatherly, Principal Investigator
Location(s): Washington
Problem
Simulation early in the system development lifecycle is becoming more prevalent than ever. Building large simulation systems requires iterative development to determine the cost and utility of system requirements. Iterative development requires small, agile teams backed by modern tools. Today scalable, efficient simulations can be built only in proprietary languages that lack tools, third-party software, and developer culture.
Objectives
We will make Java, with all its industry support and culture, a language suitable for building significant military and civilian simulations. Specifically, we will add support to Java for efficient, scalable process-based simulation.
Activities
We will equip Java to support efficient, scalable co-routines and create an application-programmer interface (API) that supports efficient simulation while retaining access to all Java features. Our approach will maximize the likelihood that the modifications and additions can be absorbed into the Java standard. We will apply the simulation API to problems in air traffic control and military simulation.
Impact
This work will enable small teams of programmers to construct significant simulations because the development tools available with Java will make programmers more productive. It will exploit Java's rich resources in graphics, databases, communications, XML, etc., and affect acquisition of large simulation systems by enabling "build a little, test a little" exploration of the cost and utility of system functions.
^TOP
Modeling and Simulation of Large-Scale Networks
Gary Comparetto, Principal Investigator
Location(s): Washington
Problem
To ensure that the Global Information Grid (GIG) can meet its needs for secure, reliable mobile communications, the DoD must model and simulate the end-to-end performance of large-scale heterogeneous communications networks, using realistic operational scenarios and traffic loads. Limitations on run time and system memory have so far made this impossible for scenarios involving more than about 100 nodes and traffic loads greater than about 10 Mbps.
Objectives
We will investigate, quantify, and document methods to improve simulation run time and memory footprint performance. Additionally, we will develop a modeling and simulation (M&S) testbed that can simulate the behavior of large-scale, heterogeneous communications networks and enable analyses of critical design and engineering issues.
Activities
This project will investigate and quantify the improvements achievable by using various simulation kernels and their parallel processing capability, co-simulation techniques and ways to share information between simulation kernels, and OPNETs high-level architecture interface. We will develop and apply an M&S testbed that incorporates our findings and demonstrate it against realistic large-scale communications network scenarios based on the DoD Transport vision.
Impact
This research effort will result in the development and demonstration of a large-scale communications network M&S testbed that exhibits the best achievable run time and system memory performance using a new combination of processes, algorithms, and technology. Timely, accurate, and effective simulation-based analyses will help the DoD to achieve its vision for the GIG.
^TOP
Multi-Organizational Human-in-the-Loop Simulation
Patti Liguori, Principal Investigator
Location(s): Washington
Problem
Joining air traffic control (ATC) simulations across organizations using the High-Level Architecture is a complex and challenging problem. The framework for this environment must be developed from the outset to meet multiple application goals that can be transferred and applied to new objectives.
Objectives
The objective of this research is to further develop distributed aviation simulation tools and standards to meet the needs of the broader aviation community.
Activities
Over the past year, this project developed an environment for performing real-time ATC simulation over the Internet. This environment, which accommodates both ATC voice and data communications, is called AviationSimNet™. MITRE is requesting organizations to help develop standards for AviationSimNet and to refine its usability to meet the needs of the broader aviation community.
Impact
A model federation for sharing flight simulators across organizations will provide opportunities for aviation organizations to pursue further exploration of air/ground concepts and may offer cost-effective methods to address air/ground research issues. It could also provide an opportunity for integrated research across organizations in support of the FAA and other government agencies.
^TOP
Next Generation Model of the National Airspace System
Brian Simmons, Principal Investigator
Location(s): Washington
Problem
Determining the system-wide impact of a local change to the National Airspace System (NAS) is a generic problem with specific instantiations. This project seeks to create a simulation tool built around a general architecture that runs flights in the NAS and allows analysts to customize that architecture to a particular problem. This tool will be used within MITRE to study the impact of changes in aviation policies/procedures, natural events, or NAS infrastructure on overall NAS performance.
Objectives
This project will enhance the tool to incorporate core functionality and simulate events that occur during the life of the flight. We will also add enhancements that allow analysts to modify and customize the model to incorporate algorithms that simulate events during the life of a flight.
Activities
This research will lead to a flexible simulation architecture that is readily changed. The approach involves developing tools that allow analysts to easily manipulate and change the simulation language and integrate the language into the overall software system. The simulation architecture and ease of reconfiguration should provide a powerful tool to analyze impacts of various changes to the NAS.
Impact
As system-wide modeling is one of MITRE's core businesses, building a tool around core functionality with the flexibility to modify events for specific scenarios will greatly improve our ability to address questions about system-wide impact, which arise frequently.
^TOP
Runway Capacity Modeling for Complex Airports
John Barrer, Principal Investigator
Location(s): Washington
Problem
The Airfield Capacity Model (ACM), which represents the current state of the art in airport capacity modeling, provides first-order estimates of the capacity of some runway systems but fails to support the analysis of many of the increasingly complex airport layouts that need to be analyzed.
Objectives
The purpose of this research is to develop a flexible airport capacity model that can handle these increasingly complex analyses. During the first year of this project we designed and implemented a prototype of this new simulation-based model, called the Airport Capacity Analysis Through Simulation, or ACATS, model. Our current objective is to confirm that the new approach (1) produces valid capacity estimates, (2) can be modified to produce delay measures, and (3) if needed, can be integrated with national and global models to provide dynamic capacity estimates within those simulations.
Activities
We will continue development of our working prototype, compare its performance to the existing ACM, and then build a working model for broader use within MITRE.
Impact
The new ACATS model overcomes the current inability to model complex airports and will improve MITRE's ability to respond to the needs of our stakeholders. The ACATS model will replace the current ACM as the tool for evaluating major investment decisions in aviation infrastructure. Improving the connections to other MITRE modeling tools will enhance our total modeling capability.
^TOP
Simulation and Modeling UAV Activities in NAS
Matt DeGarmo, Principal Investigator
Location(s): Washington
Problem
Unmanned Aerial Vehicles (UAVs) being proposed for operations in the National Airspace System (NAS) have different performance characteristics than manned aircraft. Further, missions planned are rarely point-to-point, but rather involve some form of patterned flight or tracking activity that may include intermittent short- or long-term orbits; these flights can last for days or even months. Accommodating these needs in the NAS poses numerous challenges.
Objectives
Simulation and modeling of potential UAV operational scenarios will assist MITRE, our sponsors, and other stakeholders in understanding how these vehicles and their missions will integrate with existing and future traffic flows, airspace, infrastructure, and air traffic procedures. Results will be useful in determining if changes are required and what, if any, unique restrictions will need to be imposed on UAV flight operations and vehicles.
Activities
This research will simulate and model UAVs in various flight environments, traffic mixes, and operational scenarios to determine impacts on the NAS. A further activity will be to visualize UAV operational concepts to facilitate understanding of their unique operating and performance characteristics and mission profiles.
Impact
This work will provide both quantitative and qualitative measures of potential UAV impacts on NAS structure, capacity and operations. This information could help the FAA in strategic planning of air traffic procedures, airspace design, controller training, and infrastructure resources. It could also be used by UAV manufacturers to improve designs, and by prospective operators/investors as a basis for understanding future operational constraints and opportunities.
^TOP
|
