SBIR/STTR Award attributes
The key to enabling successful Naval aviation operations is safe, trained, and ready personnel. Current approaches to improving aviator performance, safety, and effectiveness have critical limitations. First, the Navy does not fully leverage the wealth of data collected and maintained in operational and training settings to make improvements. In part, this is because harnessing the data to gain knowledge and understanding is no easy task. The complexity, scale, and interdependence of modern training and operations exacerbate well-established big data challenges, particularly the volume, velocity, variety, and veracity of the data generated. Another key challenge is to correlate and fuse multiple, often disparate, data sources to provide a complete understanding of operational and training needs, readiness, and even the impact of doctrine. For example, some problems may not be fully understood unless one looks at both aircraft and aircrew data versus looking at either data sources alone. Mapping controlled training environment data to noisy operational data makes this even more challenging. Second, the challenge of big data is not solely an engineering challenge of moving and combining data, but also of identifying patterns and relationships that focus on value creation and realization. When data are leveraged, they are typically used in a descriptive, rather than a predictive or prescriptive, form. Predictive and prescriptive analytics are used widely within industries such as marketing and sales, transportation, and oil and gas, as well as in financial markets, to optimize operations, grow sales, control prices, and manage risk. Within the context of Navy training, predictive and prescriptive analytics can help staff training pipelines, select and assign trainees to an optimal training schedule, and optimize technology selection such as when to use live versus virtual or augmented training. Through predictive and prescriptive analytics, both cost and risk of loss of life can be possibly mitigated. Further we may be able to explore the transfer of one kind of training to a different future training or operations environment. We may also be able to predict the interaction between the training medium and the timing of delivery. To meet these challenges, the Aptima Team, along with the University of Memphis and InnovaSystems, propose to develop PARETO: Predictive Analytics to Realize Effective Training and Operations. PARETO will leverage aircraft and aircrew data to make predictions and recommendations in support of operational and training effectiveness. PARETO will consist of four critical components: (1) a layered data platform, (2) a prediction engine, (3) a recommendation engine, and (4) a user interface.
