Northstrats proposal for N231-009 - DIGITAL ENGINEERING - High-Speed Data Return for Tactical Environments details Northstrats approach to exploit data collected by U.S. Navy air platforms and sensors and deliver actionable conclusions to tactical decision-makers in a mission-relevant timeframe. Northstrat will investigate and design a prototype solution that can collect and aggregate sensor data for fast, dynamic, and intuitive analysis. The solution must operate in near real-time, handle high data volumes, and be scalable to include many data types and interfaces. Various machine learning (ML) techniques will be considered to exploit the data and provide desired outcomes to aircrews, mission planning cells, and others. ML approaches can exploit data in different ways, including predictive analysis, feature recognition, or classification. ML methodologies help to take the existing long-loop manual process out of the mission-critical timeframe and replace it with real time ML inferences that provide the relevant tactical information required for the mission. During Phase I, Northstrat will address the following focus areas in order to gain a better understanding of the data exploitation challenges faced by NAVAIR. 1. Data Sources & Types 2. ML Training data 3. Data Access 4. Implementation Environment 5. Machine Learning 6. End Users 7. Requirements 8. Data Infrastructure Northstrats user engagement process enables rapid discovery of user needs, relevant use cases, constraints, and challenges. We believe the most valuable systems are designed by integrating end users into the development process. During Phase I, Northstrat will plan interviews with various NAVAIR personnel. These interviews are used to identify the use cases, available sensor data, and the desired outcomes that are of the highest value to NAVAIR users. Different problems require different ML techniques to solve. Phase I efforts will use relevant missions and desired outcomes provided by users to select ML approaches that are appropriate for each case. Descriptive, diagnostic, predictive, and prescriptive analysis processes will ensure all possible factors and scenarios inform actionable decision making. Northstrat is well positioned to address the data exploitation challenges faced by NAVAIR. Northstrat has provided software and systems engineering services to DoD and IC customers for more than 16 years. In this time, Northstrat has consistently delivered mission critical systems from inception to operations. Developing systems for the DON, NRO, USAF, USSF, NGA, FBI, DTRA, and others, we specialize in providing solutions rooted in a deep understanding of our customers problem space. We have deployed operational data analytics solutions built on both industry-leading commercial platforms and custom code. The key to our success is our highly skilled workforce and their mission-first approach to software design, integration, and development.
Benefit: The primary benefit of this project is that better data exploitation can lead to better outcomes for the U.S. Navy and the military as a whole. Access to better information in the mission-relevant timeframe means tactical decision-makers can make better, more well-informed decisions that will keep men and women in uniform safe from harm and increase the probability of mission success across the board. At the end of Phase I, Northstrat expects to have a solid understanding of the problem and the desired outcomes and use cases for NAVAIR personnel. We will have identified the primary use cases, sensor data sources, and machine learning (ML) models required to provide rapid actionable data to tactical decision-makers. Phase II will incorporate the knowledge gained in Phase I to build a prototype. This prototype will use actual sensor data for model training and demonstration of the inference engine. Northstrats approach to developing large systems relies on building an initial system that handles small, targeted use cases. These use cases work end-to-end and allow our users to realize operational mission value very early on. Northstrat then builds on that initial system to scale to more complex use cases. Our fail-fast approach to system development allows us to eliminate approaches that dont work without spending significant investment to do so. Concepts that show early promise are matured into operationally viable solutions.
Keywords: ETL, ETL, data exploitation, user engagement, data analysis, Machine Learning, DevSecOps, data aggregation, inference engine
