H9240523P0012

Small Technology Transfer Research Program (STTR) Phase I

AI/ML Aided Aviation Sensors for Cognitive and Decision Optimization

AI/ML Aided Aviation Sensors for Cognitive and Decision Optimization

Sensor systems aboard aircrafts address unique problems and are siloed in their objectives. A data silo is a term used to describe a data system that is insulated from other data systems. While keeping information categorized may lead to easier organization, the costs often outweigh the benefits. In aviation systems, data silos often lead to miscommunication, cognitive overload, and waste. These data silos also impede the ability to better optimize existing sensors on the aircraft to support other applications. Because onboard sensor capabilities are insulated, the sensor systems nor data available are not comprehensively used together beyond their primary purpose to provide advanced insights for aviators. Data science, and machine learning in particular, is rapidly transforming scientific and industrial landscapes. The defense aerospace industry is poised to capitalize on big data and machine learning, which excels at solving the types of multi-objective, constrained optimization and augmented decision-making tools. Indeed, emerging methods in machine learning may be thought of as data-driven performance optimization techniques that are ideal for high-dimensional, nonconvex, and constrained, multi-objective optimization problems, and that improve with increasing volumes of data. The data generated from aircraft siloed sensor systems can be fused together to provide higher quality and structured data to the aviators. Specifically, data obtained from multiple sensor systems can be fed into machine learning (ML) models that can provide more structured and useful insights to aviators. ML is a growing set of optimization and regression techniques to build models from data. There are a number of important dichotomies with which we may organize the variety of ML algorithms. These include verification of individual sensor data, detection of anomalies that would otherwise not be sensed, and construction of a clearer sight picture. Furthermore, these models can be trained on data collected during missions for continuous improvement of model inference. This opportunity represents an important step forward in unifying sensor system data onboard aircraft and enabling a new wave of capabilities to increase lethality, safety, and mission effectiveness while leveraging the existing hardware infrastructure onboard the aircraft.

The goal of phase I is to establish the technical merit, feasibility, and commercial potential of proposed R&D efforts and determine the quality of performance of the small business awardee organization. STTRs are completed in conjunction with a research institution.

Johns Hopkins University Applied Physics Laboratory

SOCOM23B-001

S23B-001-0048

23.B

Adriana Avakian