Current Projects

Learn more about DI's capabilities :

View completed projects

Training Systems Design and Evaluation

Virtual Technologies and Environments (VIRTE)

Under the Office of Naval Research's VIRTE program’s Demo II, the DI team has developed models and theories for multimodal sensory integration in support of training transfer, a human performance metric toolkit that can be used to assess training transfer, and a set of training transfer studies that aim to validate the theorized models, theories, and metrics that are relevant for the evaluation of VIRTE Demo II systems. The goal of this effort is to develop design science to enhance reciprocity between training theory and practice. Under VIRTE Demo III efforts, DI developed system requirements, metrics, and scenario manipulation variables for the Multi-platform Operational Team Training Immersive Virtual Environment (MOT2IVE), a deployable USMC Fire Support Team Trainer. DI has also developed a Performance Assessment and diagnostic Tool (PAST), to assist instructors and deployed troops in effectively diagnosing and interpreting performance. Additionally, the DI team has also conducted evaluations for Multi-purpose Supporting Arms Trainer (MSAT) and the MOT2IVE) System, including cue fidelity evaluations, training effectiveness evaluations (TEEs), and usability evaluations.

Online Team Performance Assessment

MAP IT-C describes a conceptual model and partially functional prototype for predicting, assessing, and analyzing team performance on line. The predictive capabilities of this tool provide trainers with ‘hotspot’ vulnerabilities with regards to team interactions and task parameters, based on theoretical evidence, team modeling hypotheses, and contextually driven operational task factors. The assessment capabilities utilize detailed task analysis metrics to assess team performance online. Analysis capabilities leverage logic to derive ‘what’ happened that caused a breakdown, in addition to ‘why’ it happened, utilizing EEG and eye tracking measures.

Founded in team performance theory, and grounded in contextual task analysis, validation studies and proof of concept prototypes are underway to illustrate the powerful predictive capabilities of MAP IT-C, over and above current state-of-the-art assessment methods for teams.

Applications of this tool range from training to operational environments. In training, the predictive capabilities allow instructors to monitor breakdowns in team coordination and pinpoint how and why they occurred, supporting directed team feedback. Operationally, this tool allows online monitoring of breakdowns, and can either utilize interface mitigation or team ‘flags’ to alert when an error with critical consequence potential occurs, supporting high levels of situation awareness, quick detection of errors, and improvements in mission safety and effectiveness.

Adaptive & Intelligent Training Environment (AITE)

Under the Office of Naval Research’s HPT&E program, DI is participating on the AITE team with UCF’s Institute for Simulation and Training. The DI team is performing a Task Analysis on the Distributed Operations (DO) concept in order to identify USMC domains in which DO are critical. From these domains, DI will derive overlapping KSAs in order to feed training objectives, system requirements and metrics to enable development of experiential learning environments to facilitate accelerated learning, enabling Marines to train like they fight.

Infantry Immersive Trainer (IIT)

Design Interactive (DI) supported the development of the Infantry Immersive Trainer (IIT) by both developing front end user requirements and by conducting a training effectiveness evaluation (TEE). To ensure that training effectiveness was built into the system from early in the design process, task analysis data were used to develop cue fidelity requirements and requirements for scenario events designed to provide opportunities to practice key tasks. After the system was fully instantiated, DI conducted a two-part evaluation of the IIT training program to examine the degree that current hardware and software installations support targeted training goals as listed in the Pre-deployment Training Package (PTP). The TEE uncovered strong value in the training system, and identified a number of redesign recommendations that, if addressed, would improve training efficacy (by up to 50% for some tasks). Key improvements focused on increasing hardware or software fidelity, or on providing additional scenario training opportunities to ensure that all targeted objectives could be met, further improving the impact of the system.

Tool for the Optimization of Multimodal Cues in Advancement of Training System Design (TOMCAT)

Ensuring that training systems are based on operationally, theoretically, and empirically driven requirements is a key component to improve VE system effectiveness as it focuses training on targeted goals. System designers may not have access to specialist resources to develop these requirements, and there is little guidance available on the process. TOMCAT aims to guide developers through the process of extracting the necessary contextual data to develop system requirements which support training effectiveness, and then allowing them to optimize these requirements by integrating theoretical, empirical and cost benefit findings.

Observational Skills Enhancement and Retainment Virtual Environment (ObSERVE) and Multi-Axis Performance Interpretation Tool (MAPIT)

DI is supporting development of the ObSERVE observational trainer, a deployable simulation aimed at increasing observational skills of infantry Marines. Observational skills are critical to situational awareness (SA) and tactical decision making. DI is performing a detailed task analysis which will drive ObSERVE system requirements to facilitate advanced observational skills training. In parallel, DI is designing the MAPIT training management system which will leverage neurophysiology technologies to support data collection, diagnosis and mitigation within the VE.

Naturalistic Decision Making (NDM)

DI is conducting task analysis on operational ad hoc teams to identify the cognitive and behavioral processes underlying team consensus and decision making for one-of-a-kind or unique task situations.

Human-System Integration

Affective Virtual Environment Training System (A-VETS)

Under an ONR Phase I SBIR effort, a conceptual framework, the Affective Virtual Environment Training System (A-VETS), was developed to guide appropriate inclusion of affective/stress cues into a virtual training environment to enhance learning and training transfer. This is particularly important for military training systems where the adaptation to negative affective factors such as the theater of war (including live fire, excessive noise, and dynamic threats that impact operations at a moment’s notice) are critical for optimal performance in the transfer environment. The A-VETS framework includes components to (a) guide instructors in selecting a domain and targeted training objective(s), (b) develop a set of guidelines to guide system designers through the integration of affect induction cues into the training environment, and (c) provide a personalized training strategy that optimizes presentation of multimodal cues in real-time to elicit appropriate emotional experiences while maintaining trainee performance levels.

Design Interactive, Inc. has been awarded a Phase II SBIR to fully develop the A-VETS tool, and has partnered with VRSonic who is developing noninvasive methods to evaluate emotional responses of trainees in real-time. The envisioned system will evaluate trainee emotional response and performance levels in real-time and dynamically adjust the degree of affect within training sessions. While the current effort targets a specific need for the military training community, the A-VETS tool is applicable to many domains that require performance under high levels of stress or extreme emotional conditions.

System for Tactile Reception of Advanced Patterns (STRAP)

Design Interactive, Inc. (DII) was awarded an ONR Phase II SBIR contract to continue development of the System for Tactile Reception of Advanced Patterns (STRAP) into a robust, lightweight tactile communication device for dismounted soldiers. During Phase I, an intuitive tactile communication library, grammar, and prototype hardware/software system was developed. Preliminary user testing has shown that the tactile language, consisting of 38 distinct symbols, can be learned in under 3 hours. During Phase II STRAP will be combined with AnthroTronix, Inc.’s Team Status and Signaling System (TS3) data capture system to create an unobtrusive bi-directional communication device referred to as the Haptic Automated Communication System (HACS).

CogGauge - Cognitive Assessment Tool

Under NASA SBIR funding, Design Interactive is developing a cognitive assessment tool for astronauts- CogGauge. Phase I work is focused on conceptual development. The tool will be in the form of a portable gaming application. CogGauge, while engaging astronauts in an entertaining experience, will use a hybrid approach to combine predictive tools for assessing cognitive workload with metrics that assess performance decrements. This approach will take into consideration learning effects and determines the best predictors of performance decrement to measure cognitive workload. CogGauge, once functioning, will detect cognitive decrements and provide immediate feedback to astronauts and/or flight surgeons regarding readiness for space operations.

Next-Generation HSI Research

Revolutionary Accelerated Processing Image Detection (RAPID) System

The objective of this effort is to enhance analyst/system effectiveness during imagery analysis by increasing both speed of image review and accuracy of threat identification. RAPID incorporates an integrated hardware and software solution for real-time capture and analysis of analyst eye and electroencephalography (EEG) data to monitor processes associated with detection and identification of relevant areas of interest when analyzing images. Eye tracking technology is being used to identify meaningful fixation points on an image, and capture oculomotor indices of cognitive load at each fixation point (e.g., fixation duration, pupil size). In addition, event-related signatures in EEG known as Fixation-locked Event-Related Potentials (FLERPs) are evaluated to determine whether the fixation was of interest to the analyst. Specific EEG classifiers are being developed to distinguish correct responses and erroneous assessments (i.e., misses or false alarms), the latter of which may then be mitigated through display adaptation strategies. The RAPID system has been integrated with complimentary services from other research projects to capture and exploit user intent during imagery review to enhance collaborative sense making.

The RAPID system has been empirically tested using static images. Stimulus-locked and response-locked EEG/ERP single trial classifiers were 79% and 95% accurate, respectively using 9 mono-EEG channels and simple feature extraction. Initial investigation into fixation-level EEG/ERP single trial classifiers is ongoing. Current studies are examining the utility of RAPID in enhancing threat detection during more natural analysis scenarios, where analysts can pan/zoom/flicker between multiple images.

This work is funded by the Intelligence Advanced Research Projects Activity (IARPA) Collaboration and Analyst System Effectiveness (CASE) Program, contract FA8750-06-C-0197 issued by Air Force Research Laboratory (AFRL).

Advanced Neurophysiology for Intelligence Text Analysis (ANITA) System

Design Interactive is currently funded by Defense Advanced Research Projects Agency (DARPA) via an SBIR Phase II to develop the ANITA system, which utilizes a neurophysiological approach of event-based assessment of operator activity. The objective of ANITA is to develop a closed-loop system that enhances the accuracy and speed of evidence gathering from text sources, and provides innovative interaction paradigms to optimize sharing of information among analysts. The ANITA system analyzes a user’s cognitive processes via eye tracking and electroencephalography event-related potentials (EEG/ERPs) during reading to support automatic relevance detection in written text. At a sentence level, ‘interest’ levels are evaluated by the physiological sensors, indicating each sentences’ relevance with regard to the users’ mental model. This physiological relevance indicator can be used to identify (and, if necessary, auto-extract) text snippets in the reviewed documents that are relevant to the current analysis goal as reflected in the user’s mental model.

The ANITA system includes three main components:

(1) ANITA Reader: Evidence gathering component of ANITA that is driven by physiological sensor hardware where text documents are displayed and relevant snippets are pulled;

(2) ANITA Shoebox: Database shared by ANITA components in order to store, add, enhance, change or delete evidence items. Innovative search and visualization features provide novel, effective approaches to organization, retrieval and discovery of individual and community-wide information;

(3) ANITA Sandbox: Evidence marshaling tool in which collected items of evidence can be organized and analyzed.

Originally developed to support the analysis process of intelligence analysts, the ANITA system is also applicable to study or research tools, training evaluations, or gaining an objective measure of “understanding” with regard to text sources.

Proactive Aiding in Command-and-Control Environments System (PACES)

To meet the challenges imposed on command-and-control environments by next-generation weapons systems and continued reduced manning efforts, Design Interactive, Inc. teamed up with Soar Technology Inc. for an Air Force SBIR Phase I to develop the Proactive Aiding in Command and Control Environments System (PACES). PACES will be an automatic agent, informed by real-time data streams from the system, the mission, and the operator's cognitive state, using dynamic constraint-based task modeling to anticipate future mission state and operator functional state (OFS) ahead of time. A component of Design Interactive’s MIDS workload analysis method will be used to calculate expected operator load for the future task demands anticipated by the model in real-time. Given this information, preventive adaptations of the information display can be dynamically applied to avoid cognitive bottlenecks before they occur. In addition to preventive adaptation, PACES will employ physiological measures, specifically electroencephalogram and eye tracking, to assess the operator's actual cognitive state and mitigate problems in real-time. Physiological measures will provide input to an intelligent Soar architecture to derive OFS indicators and inform PACES when adaptive aiding is needed. The model may further analyze workflow history and operator's physiological and behavioral responses to system events in order to dynamically adjust and improve the predictive modeling component and mitigation strategies.