You are here

Data tools for the Army Basic Training Environment

Description:

TECHNOLOGY AREA(S): Human Systems 

OBJECTIVE: Develop a service-oriented architecture that permits the use of measures from unobtrusive COTS sensors, training data, and other measures of health and wellbeing to understand, manage, and optimize the wellbeing and performance of Army enlistees in an initial entry training environment. 

DESCRIPTION: The proliferation of low cost fitness sensors provides opportunities for individuals to experiment with diet, exercise, and lifestyle to optimize key indicators of health. These devices typically link to a smartphone and allow individuals to log their own data and, at least in theory, track improvements in fitness over time. By tracking exercise, diet and fluid intake, sleep cycles, and other behaviors, users can create fitness plans and receive automatic notifications and incentives to follow that plan. Over time, it is expected that better and more diverse COTS sensors will be available and will potentially have utility for Army training environments. Current commercial fitness devices provide feedback, goal monitoring, and a range of other services to the individual consumer (Sullivan & Lachman, 2016). While some devices have open standard protocols, others are closed proprietary systems (Gay & Leijdekker, 2015; Nijeweme-d'Hollosy, van Velsen, Huygens, & Hermens, 2015). Research into strategies to change fitness behavior has shown that factors such as goal setting, feedback and rewards, coaching, and social factors are all potential avenues for effecting change. Commercial fitness trackers typically employ several of these strategies to facilitate increases healthy behavior; however, research into the benefit of fitness devices is still very new and the evidence supporting their effectiveness is somewhat mixed (Sullivan & Lachman, 2016). Perhaps in part because privacy concerns, there is not currently a market for services that aggregate fitness data across individuals with the goal of managing health and fitness at a group level. While the individual fitness market may not support this need, a training organization like the Army could potentially reap a huge benefit from this capability. Managing the physical fitness of service members has always been a core Army mission that is integral to unit readiness. The management of fitness, health, wellness, and training performance is no more important than in the basic training environment, one that is unique in a Soldier’s career. During basic training, enlistees eat, sleep, train, and live under the close supervision of their leaders on an Army post. Trainee data from fitness sensors, training events, and other behavioral and psychological measures are critical for instructors and leaders overseeing this key period of training. What is needed is an architecture to easily collect and aggregate that data across groups, to analyze, visualize, and understand it, and to effectively use it to manage outcomes by providing tailored feedback to each individual (TRADOC PAM 525-8-2). For example, suppose data revealed that consumption of high-fat foods, poor sleep patterns, and long heart rate recovery times following physical activity predict a higher likelihood of a failing Army Physical Fitness Test (APFT) score. Once this relationship is discovered, a number of interventions would be possible to behaviorally alter some of the predictors using proven methods such as reward, coaching, and goal-setting. Over time, machine learning techniques could be applied to identify which strategies are most effective at attenuating this risk. Proposals should describe your approach for designing and developing an open-standard, service oriented architecture for aggregating data from COTS sensors, training events, and other measures of health and wellbeing, and providing access to those data by tools to mine the data to discover associations among measures, and tools for designing, delivering and evaluating interventions to attempt to accentuate positive outcomes and attenuate negative outcomes. The system should provide a plug and play capability both for input devices and for analytical and intervention tools. Finally, the system should provide protocols to facilitate things like data entry, quality control and security. 

PHASE I: Determine the feasibility/approach for developing an open standard architecture for aggregating data across individuals from COTS sensors, training events, etc., so that they can be used by trainees, instructors, and leaders for understanding the relationships among measures and for designing and evaluating interventions such as personalized get-well plans. Work in this phase should include a user needs analysis to become familiar with the basic training environment and the instructors, leaders, and course managers who are involved in delivering the training. The government will insure access to the necessary user groups for this analysis. This analysis will also help the vendor to improve strategies for reducing technical risk. The phase 1 deliverable will be a design to establish the technical merit, feasibility, and commercial potential of the proposed R&D effort. The design and associated feasibility analysis should demonstrate support for the following capabilities: 1. Open standard service oriented architecture: The core architecture should enable the collection and storage of sensor and other data into a non-proprietary, open-standard format such as the experience application programming interface (xAPI) standard in use by the DoD. Additionally, the core architecture should enable third-party developers to create a variety of tools for data collection, analysis, and visualization as well as tools for developing, creating, and evaluating interventions for unit members. 2. Data collection tools and processes: An important goal of the research in the SBIR is to identify a set of potential measures and to analyze the feasibility of collecting those measures using available COTS devices. Measures may include those typically found on fitness devices as well as psychometric measures and other verbal report measures that might be collected on a mobile device. Finally a means of incorporating key training performance metrics will need to be evaluated. 3. Data mining tools and processes: Users will not have a background in data analysis and so tools need to be developed that automatically analyze and present data using visualizations that are intuitive and that address the questions that those users are most likely to have. The user needs analysis will be critical in determining the user requirements/use cases for the proposed data mining tool(s). 4. Intervention tools and processes: When relationships are found that predict good or poor outcomes (e.g., improved/worsening PT scores), intervention tools will be needed to implement behavioral modification programs to improve the likelihood of desired outcomes. Interventions should be based on proven methods of behavioral change and should also automatically assess their effectiveness. For this proposal the vendor should focus on the following outcomes: PT scores and Record Fire scores. Desirable outcomes would be improvements in performance. 5. Data integrity: Processes, technologies, and tools are needed to insure data integrity. Data integrity may be compromised by a range of issues including faulty sensors and human error. Detection and correction of data errors is an essential capability and the feasibility analysis should address how to best mitigate errors in data sets. 6. Ability to function in a training environment: The basic training environment includes everything from classroom training to field training. Training sites may have limited or no access to cellular networks and/or power supplies (for re-charging batteries). Trainees crawl, walk, and run through various types of terrain in all manner of weather at daytime and night. The analysis and design solution should address any consequences or limitations created by the training environment. 7. Intuitive user interface: As already mentioned, the user needs analysis should feed the design of the user interface. The technology solution will succeed or fail based on the design of the user interface. A system that adds to instructor workload will not be accepted by users. The user interface must insure that the benefit of the system far outweighs the cost from the user point of view. 

PHASE II: This phase will consist of the development, demonstration, and delivery of a working prototype. It is expected that an iterative design and development of components of the system will be needed. To insure good acceptance by the user community, the government will insure that the necessary users are available for evaluation of prototype interfaces etc. The Army’s IRB will need to approve any human subjects research. To facilitate approval, no PII needs to be collected for the demonstration. Determining the potential for this system to be commercially viable requires that the system’s ability to deliver the seven capabilities described above (see phase 1) be adequately demonstrated. In this phase the vendor will have to provide a plan for demonstrating each of these capabilities along with criteria for success or failure for approval by the government. Given the time frame, it will probably not be possible to demonstrate the effectiveness of interventions. The “operational” environment in this case is the basic training environment. Participants will be available as needed for this demonstration. Phase II deliverables include full system design and specifications to include executable and source code. It is expected that the final deliverable will be at a technology readiness level (TRL) 6 (System/ subsystem model or prototype demonstration in a relevant environment). As this prototype is a software architecture utilizing COTS hardware, achieving TRL 6 demonstration should be feasible. 

PHASE III: Follow on activities are expected to be aggressively pursued by the offeror to seek opportunities to integrate the hardware, software, and protocols into Army personnel and training management systems. Commercial benefits include applications of the same capability in private businesses that have wellness programs for their employees as well as to expand and apply these capabilities outside of the basic training environment in the Army. 

REFERENCES: 

1: Gay, V., & Leijdekker, P. (2015, Nov). Bringing health and fitness data together for connected health care: Mobile apps as enablers of interoperability. Journal of Medical Internet Research, 17(11). Doi 10.2196/jmir.5094. Retrieved from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4704968/

2:  Nijeweme-d'Hollosy, W.O. van Velsen, L., Huygens, M., & Hermens, H. (2015). Requirements for and barriers towards interoperable eHealth technology in primary care. IEEE Internet Computing

3: 19(4),10–19. doi: 10.1109/MIC.2015.53.

4:  Sullivan, A.N., & Lachman, M.E. (2016). Behavior change with fitness technology in sedentary adults: A review of the evidence for increasing physical activity. Frontiers in Pulblic Health, 4, 1-16. doi: 10.3389/fpubh.2016.00289. Retrieved from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5225122/pdf/fpubh-04-00289.pdf

5:  U.S. Army (2017). The U.S. Army Learning Concept for Training and Education: 2020-2040. TRADOC Pamphlet 525-8-2. Retrieved from: http://www.tradoc.army.mil/tpubs/pams/tp525-8-2.pdf

KEYWORDS: Data Analytics, Data Visualization, Data Mining, Machine Learning, Basic Combat Training, Fitness Tracking, Comprehensive Soldier Fitness 

CONTACT(S): 

Gregory Goodwin 

(407) 384-3987 

gregory.a.goodwin6.civ@mail.mil 

US Flag An Official Website of the United States Government