RESEARCH AND DEVELOPMENT

High Resolution Environment Information for Urban Operations.. 83

An Automated Short-Free-Text Scoring System:  Development and Assessment.. 84

CPU Architectures & Network Protocol Offload in a Simulation Environment.. 84

Automatic Performance Evaluation and Lessons Learned (APELL) for MOUT Training   85

Calculating Error Tradeoffs in Weapon Simulation for Live Training.. 85

Knowledge Tracing and Prediction of Future Trainee Performance. 86

A Framework for Developing Serious Games to meet Learner Needs.. 86

Human Behavior Model Interoperability For Training In Virtual Worlds.. 87

Videogame play and the effectiveness of virtual environments for training.. 87

Distributed Scheduling for Wireless Mesh Networks in Realistic RF Environments.. 88

Successful Deployment of Mesh Routers for Trans-Continental Urban Resolve Experiments   88

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

High Resolution Environment Information for Urban Operations

2006 Paper No. 2482

 

Simon Ahlberg, Dr. Ulf Söderman

 Swedish Defence Research Agency, FOI

Linkoping, Sweden

 

Many modern operations, civilian as well as military, including relief work, peacekeeping and peace enforcement, take place in urban areas, which are known to be complex and dynamic environments. Therefore, access to updated and detailed geographic (terrain) information and/or 3D-models of the operating environment is often a key factor for mission success. This kind of information is, however, often sparse, old or even unavailable. To address this problem, especially in international operations, a combination of methods and procedures for data acquisition and data preparation will most likely be used. These include downloading available data, purchasing commercial data (maps, satellite images) or acquiring new data using (surveillance) sensors followed by extracting geographic and geospatial intelligence data. 

 

This paper addresses the problems with supplying updated and detailed geographic (terrain) information and/or 3D-models for urban operations. We will present ongoing work at the Swedish Defence Research Agency (FOI) dealing with these problems. One focus is the development of new methods for processing high resolution sensor data used for rapid and automatic environment modelling for 3D visualisation, recognised environment picture, decision support, mission planning, mission rehearsal, etc. The methods include ground and feature characterisation, extraction and modelling. The long term objective is an automatic production chain from environment data collection to geospecific environment models. A joint project with the Swedish Army Combat School aimed at developing an extensive high resolution reference geographic database for the city of Norrköping will also be discussed. This project is referred to as DigNor (Digital Norrköping). It contains many data types, e.g. high resolution infrastructure information, an elevation model with 0.25m post spacing, geospecific building models and properties of individual trees such as position, height and size. …

 

This paper is available on the 2006 I/ITSEC CD ROM.

Order it from I/ITSEC'S Website.

 

 

An Automated Short-Free-Text Scoring System:  Development and Assessment

2006 Paper No. 2675

 

Ohad Bukai, Robert Pokorny, and Jacqueline Haynes

Intelligent Automation, Inc.

Rockville, MD.

Computerized evaluation of students’ free-text answers is already used in complex, high stakes testing, such as some of those conducted for advanced placement examinations.  These automated scoring systems work best when student answers contain 200 or more words, and when the evaluation system has many examples, exemplifying both good and poor answers. We developed and assessed an algorithm which evaluates student short answers when the answer has as few as 10 words, with the answer key using only one correct answer.  We sought such an algorithm so that individual instructors could ask students to produce short open-ended text responses to questions.  Our algorithm automates the scoring of free-text answers, enabling instructors to embed such questions in online courses, and providing nearly immediate scoring and feedback on student responses.  This algorithm is based on the semantic relatedness of the words in the student answer to the single correct answer provided by the instructor.  The semantic relatedness algorithm requires a dedicated domain specific index or collection of topic-focused documents (a corpus), which is created by an automated crawl mechanism that collects documents based upon descriptive domain keywords.    

 

We assessed the accuracy of this algorithm by collecting student answers to two text questions about botany. This material was not complex, but required students to understand what they read about Dutch Elm disease. Sixty-three students read the material, and submitted answers to two questions.  Across two questions, students’ answers averaged 9.5 words; the model answers from the instructor averaged 20 words. Across the two questions, the algorithm’s scores correlated .76 with the scores from a panel of four human raters. For the same answers, individual human raters’ ratings correlated .88 with the other three raters.  While the algorithm was not as accurate as individual human raters, we believe it shows enough promise to investigate in applied field tests with real instructors and students. 

 

This paper is available on the 2006 I/ITSEC CD ROM.

Order it from I/ITSEC'S Website.

 

 

 

 

CPU Architectures & Network Protocol Offload in a Simulation Environment

2006 Paper No. 2805

 

Jeremy Buboltz, Jacob Engel, Taskin Kocak

School of Electrical Engineering and Computer Science

University of Central Florida

Orlando, FL

 

Matthew Jemtrud

Northrop Grumman

Orlando, FL

 

The simulation industry is increasingly using server CPUs for processing data.  Simulation data creates a large number of small packets that must be processed at the network level before the higher level simulation processing is performed. As network traffic increases it becomes a burden on the server CPUs, which need to process more packets and still have room for application software. Dual-Core processor systems are being introduced as a solution to the increase in processor utilization. In this work, we present a performance analysis for servers based on the dual-core AMD Opteron and the dual-core Intel Xeon processors while processing in a typical simulation environment. Test log files were generated for a representative packet clusters. These files were then used to compare Intel dual-core and AMD dual-core servers verse a baseline dual-processor server using both unidirectional and bidirectional communications.

 

This paper is available on the 2006 I/ITSEC CD ROM.

Order it from I/ITSEC'S Website.

 

 

 

Automatic Performance Evaluation and Lessons Learned (APELL) for MOUT Training

2006 Paper No. 2494

 

Hui Cheng,  Rakesh Kumar, Thomas Germano and Chris Meng

Sarnoff Corporation

Princeton, NJ, USA

 

In order to train war fighters for urban warfare, live exercises are held at various MOUT facilities. Commanders and instructors need to have situation awareness of the entire mock battlefield, and also the individual actions of the various war fighters. The commanders and instructors must be able to provide instant feedback and play through different actions and ‘what-if’ scenarios with the war fighters. The warfighters in turn should be able to review their actions and rehearse various maneuvers. The system must be able to automatically score the performance of the warfighters and provide them an assessment of their performance. 

 

In this paper, we describe the technologies behind a prototype training system, which tracks and automatically assesses performance of war fighters around an urban site using a combination of ultra-wideband RFID, INS pose, trigger sensor and smart video based tracking.  The system is able to: 1. Tag each individual with a unique ID using an RFID system. 2. Track and locate an individual’s position, head pose and weapon pose within the domain of interest at all times during an exercise. 3. Associate IDs with visual appearance derived from live videos.  4. Visualize movement and actions of individuals within the context of a 3D model. 5. Store and review activities with (x, y, head pose, weapon pose, gun-trigger, ID) information associated with each individual. 6. Automatically create events and performance metrics for each warfighter. These events are stored in a database. User can click on an event and see the associated video. 

 

An ontology is used to represent the expert knowledge for MOUT training. Using this ontology, the tracks are analyzed, performance metrics and events are automatically created. The metrics of a novice can be compared…

 

This paper is available on the 2006 I/ITSEC CD ROM.

Order it from I/ITSEC'S Website.

 

 

Calculating Error Tradeoffs in Weapon Simulation for Live Training

2006 Paper No. 2515

 

Robert J. Hall

AT&T Labs Research

Florham Park, NJ

 

Matthew R. Janisz

Applied Research Associates, Inc.

Orlando, FL

 

Next generation instrumentation systems for live military training will simulate weapon engagements (shots) using geo-pairing: matching shooter with target by geometric computations that determine whether the trajectory or blast of the round intersects the target. Such calculations depend critically on accurate sensor readings, including positions of shooter and target(s), the pointing angle of the weapon at the time of the simulated shot, and positions of terrain obstacles. However, perfect sensors are unattainable, so a key question for system engineers is just how accurate do the sensors need to be? After all, physical weapons themselves have inherent inaccuracies (weapon “dispersion” or “spread”); for training purposes, it is not necessary to simulate better accuracy than exists in the weapons themselves. Moreover, a less accurate point angle sensor can to some degree be compensated for by a more accurate position sensor, and vice versa. Thus, ultimately we would like to understand this trade off quantitatively in order to support cost effective system engineering decisions.

 

This paper describes an iterative statistical approximation method, implemented as a set of computational tools, used by the U.S. Army’s One Tactical Engagement Simulation System (OneTESS) Project for supporting these design trade-offs. In particular, we show how to compute the probability of correct sensor-based shot adjudication forgiven sensor combinations, as well as trade off diagrams showing which combinations of sensors can be used together to achieve realism matching the physical accuracy of actual weapons systems. …

 

This paper is available on the 2006 I/ITSEC CD ROM.

Order it from I/ITSEC'S Website.

 

 

 

Knowledge Tracing and Prediction of Future Trainee Performance

2006 Paper No. 2992

 

Tiffany S. Jastrzembski

Florida State University

Tallahassee, Florida

 

Kevin A. Gluck,  Glenn Gunzelmann

Air Force Research Laboratory

Mesa, AZ

 

Intelligent tutoring systems seek to optimize instruction and training by adapting and individualizing the learning experience on the basis of a student model (Shute, 1995). This model represents the system’s estimate of the student’s current knowledge or skill level, established from a performance history. Knowledge tracing (Aleven & Koedinger, 2002; Anderson, Conrad, & Corbett, 1989) is a dynamic, Bayesian approach to updating the estimates of probability of skill mastery in the student model. A fundamental shortcoming of this approach is that it does not include a representation of memory decay during periods of non-practice. As a result, traditional student modeling approaches are unable to make predictions regarding knowledge and skill changes under various future training schedules or to prescribe how much training will be required to achieve specific levels of readiness at a specific future time. In this paper, we propose a new knowledge tracing equation, computationally inspired by the learning and forgetting equations in the ACT-R cognitive architecture (Anderson et al., 2004), which uses performance history to baseline student model parameters and then extrapolates knowledge state transformation to predict future performance. We explore practical issues concerning predictive models of future trainee performance and the prescription of frequency and timing of optimal learning with training systems. For instance, we investigate how much data from the training history are necessary to achieve reasonable predictive validity, and we describe the impact of data granularity through a quantitative assessment of how adequately the model can fit and predict human performance curves across aggregate-level, team-level, and individual-level resolutions. …

 

This paper is available on the 2006 I/ITSEC CD ROM.

Order it from I/ITSEC'S Website.

 

 

A Framework for Developing Serious Games to meet Learner Needs

2006 Paper No. 2742

 

Sara de Freitas PhD

London Knowledge Lab, University of London

London, England

 

Steve Jarvis

Vega Group PLC

Welwyn Garden City, England

 

This paper will communicate preliminary findings from applied research exploring how to ensure that serious games are cost effective and engaging components of future training solutions. The applied research is part of a multi-million pound program for the Department of Trade and Industry, and involves a partnership between UK industry and academia to determine how bespoke serious games should be used to best satisfy learning needs in a range of contexts. The main objective of this project is to produce a minimum of three serious games prototypes for clients from different sectors (e.g., military, medical and business) each prototype addressing a learning need or learning outcome that helps solve a priority business problem or fulfill a specific training need. 

 

This paper will describe a development process that aims to encompass learner specifics and targeted learning outcomes in order to ensure that the serious game is successful. A framework for describing game-based learning scenarios is introduced, and an approach to the analysis that effectively profiles the learner within the learner group with respect to game-based learning is outlined. The proposed solution also takes account of relevant findings from serious games research on particular learner groups that might support the selection and specification of a game. A case study on infection control will be used to show how this approach to the analysis is being applied for a healthcare issue.

 

This paper is available on the 2006 I/ITSEC CD ROM.

Order it from I/ITSEC'S Website.

 

 

Human Behavior Model Interoperability For Training In Virtual Worlds

2006 Paper No. 2838

 

David McDonald,  Richard Lazarus, William Ferguson,  Alice Leung, Talib Hussain

BBN Technologies

Cambridge, MA

 

Barry Silverman

University of Pennsylvania

Philadelphia, PA

 

It is commonly recognized that a key challenge of urban military operations is the complex terrain. However, the large number and diversity of people present another less recognized but critical challenge, especially for non-combat operations. Warfighters must understand the socio-cultural effects of their actions, handle direct interaction with non-combatants (and potential combatants), and understand the immediate and long-term consequences of their interactions: every soldier must be a skilled ambassador. Given that realistic training requires many participants, the high cost and limited availability of highly-qualified participants restricts access to quality training opportunities. Synthetic characters can provide the numbers, but present-day simulated characters have limited capabilities for interacting with each other and with humans in the rich, meaningful way that is required. Thus, we must populate these virtual training worlds with characters whose behaviors are generated by human behavior models (HBMs) rather than humans. While significant achievements have been made in developing HBMs that are able to control a single simulated character (or a single group of simulated characters), a serious limiting factor has been the inability of heterogeneous HBMs to interact with each other or share knowledge about their environment. We present an architecture and multi-level message framework for enabling HBMs to communicate with each other about their actions and their intentions. Also, we describe our conception of the “Mindscape,” which will facilitate the use of massive numbers of synthetic characters by representing the options and state of social relationships and interactions as shared affordances in the environment. ...

This paper is available on the 2006 I/ITSEC CD ROM.

Order it from I/ITSEC'S Website.

 

 

 

Videogame play and the effectiveness of virtual environments for training

2006 Paper No. 3001

 

Rebecca Tortell,  Jacquelyn Ford Morie

University of Southern California Institute for Creative Technologies

Marina del Rey, CA

 

The Sensory Environments Evaluation (SEE) project set out to examine the effects of emotional valence of a virtual training scenario on learning and memory. Emotional arousal is well-established as having enhancing effects on memory (McGaugh, 2000). A virtual scenario called DarkCon was created to resemble a night-time reconnaissance mission. Priming of subjects was the first experimental variable. Subjects were randomly assigned to receive their mission briefing in a serious style, suggesting a serious military mission, or in a lighter style, suggesting a fun role-playing game. The influence of videogame experience was included in analysis of subjects’ recall of the environment and of their physiology. In the present study, 34 Army Rangers from Fort Benning, GA underwent the DarkCon mission. Significant effects of priming condition and videogame play were discovered in subjects’ recollection of the mission, and in their physiological reactions to highly exciting material. This paper is primarily concerned with the effects of videogame play frequency on subjects’ behavior, recall, and physiology. The effects of priming will be cursorily discussed here as they relate to videogame play habits, and explored in more detail on their own in future publications. Directions for future research into the effects of videogame play experience on training are discussed.

 

This paper is available on the 2006 I/ITSEC CD ROM.

Order it from I/ITSEC'S Website.

 

 

 

Distributed Scheduling for Wireless Mesh Networks in Realistic RF Environments

2006 Paper No. 2635

 

Hua Zhu,  David Tang,  Bo Ryu

San Diego Research Center, Inc.

San Diego, CA

 

Ajay Gummalla

Member, IEEE

 

In this paper, we evaluate the performance of two distributed scheduling algorithms for wireless mesh network through simulations with realistic RF modeling. 

 

Wireless mesh networking has recently attracted lots of interests in both academia and industry, and becomes a critical technology in pursuit of the 21st century joint force operations. One major challenge in mesh networks is how to maximize the network capacity, which is strictly limited by interference. Two distinct distributed scheduling algorithms have been proposed for IEEE802.16/WiMax and OT-TES (Operational Test-Tactical Engagement System being designed for the U.S. Army by SDRC), respectively. Despite sharing the same goal of achieving efficient collision-free Time Division Multiple Access (TDMA) scheduling, they are based on very different paradigms of distributed node coordination. The 802.16 approach, namely Coordinated Distributed Scheduling (CDS), is based on simplified quasi-interference model in which wireless networks are approximated with wired-like logical connectivity graphs. In contrast, the OT-TES approach based on Link State Protocol (LSP) does not rely on such RF modeling approximation, and instead fully embraces realistic non-quasi interference environments. As a result, in a realistic RF environment, 802.16’s CDS algorithm may fail to achieve collision-free transmissions while SDRC’s LSP approach is truly collision-free.  

 

This paper is available on the 2006 I/ITSEC CD ROM.

Order it from I/ITSEC'S Website.

 

 

 

Successful Deployment of Mesh Routers for Trans-Continental Urban Resolve Experiments

2006 Paper No. 2569

 

Thomas D. Gottschalk

CACR, Caltech

Pasadena, California

 

Robert Lucas, Gene Wagenbreth,  and John Tran

Information Sciences Inst. /Univ of So Calif

Marina del Rey, California

 

William Helfinstine

Lockheed Martin Info

Burlington, Mass

 

This paper reports the successful deployment of a robust scalable interest-managed router architecture that has supported a series of trans-continental simulations, such as Urban Resolve. Previous architectures had served well over the years, but were conceptually limited both in scalability and in robustness, or fault-tolerance. The scalable router architecture had its inception in high performance parallel computing research and its initial application in a truly scalable architecture for inter-node communications on parallel supercomputers and Linux clusters. Its design provided both needed scalability and desirable robustness on the single platform meshes of several large parallel computers made up of hundreds of compute nodes. The scalable router was designed to integrate smoothly with other Urban Resolve software by reusing Run Time Infrastructures (RTI-s) components. In an effort to minimize communication latency, maximize use of available network bandwidth, and increase robustness of trans-continental (Virginia to Hawaii) operations, Joint Forces Command’s J9 directed that its wide-area router’s offer the same characteristics of scalable and robust operations. That led to the wide-area deployment …

 

This paper is available on the 2006 I/ITSEC CD ROM.

Order it from I/ITSEC'S Website.