Data Science Driven Aircrew Performance Measurement and Proficiency System
Navy SBIR FY2018.1

Sol No.: Navy SBIR FY2018.1
Topic No.: N181-026
Topic Title: Data Science Driven Aircrew Performance Measurement and Proficiency System
Proposal No.: N181-026-0054
Firm: Perceptronics Solutions, Inc.
3527 Beverly Glen Blvd.
Sherman Oaks, California 91423
Contact: Robert Jacobs
Phone: (818) 991-8455
Web Site:
Abstract: This proposal is to develop a Data Science Driven Aircrew Performance Measurement and Proficiency System (AviatorDX). AviatorDX will integrate data collection and fusion of metrics from multiple modalities in order to provide a set of comprehensive performance measures for Live, Virtual and Constructive flight training exercises. Our approach is to adapt and exploit advanced data science approaches, including Bayesian inference technology, to develop an optimal data solicitation schema that will improve its competency estimation model through machine learning as it gains experience. Perceptronics Solutions recognizes the criticality of performance assessment as an element of effective training. For this project, we will leverage two directly related previous Navy efforts; these are: (1) MPMS (Multi-Modal Performance Measurement System) is an assessment tool that will capture performance data from a variety of live, virtual, and constructive sources. MPMS will also provide data-derived performance measures, identify skill deficits, trace deficit sources, and prescribe remediative performance enhancing evolutions. (2) DECIDE (Diagnostic Engine for Cognitive Improvement and Decision Effectiveness) will provide AviatorDX with an innovative methodology for observer-based training assessment, diagnosis and improvement. We will adapt and extend the capabilities of DECIDE and MPMS to meet the goals of the AviatorDX project.
Benefits: Our proposed AviatorDX solution offers software architecture and related software modules that will provide a capable, yet flexible, approach for collaborative data collection, shaping, fusion, and visualization. It will allow users to capture training data from a variety of sources such as physiological sensors, simulation software/hardware, aircraft avionics, as well as subjective observations of a trainer, and utilize a set of fully customizable performance measures to shape or fuse this data. The resulting measures will automatically: (1) Provide customized insight to the current level of demonstrated performance, (2) Track performance trends over time, (3) Diagnose performance deficit origins: and (4) Assess training progress and proficiency in relation to exit criteria. AviatorDX will also provide the ability to draw conclusions based on incomplete data with the help of probabilistic reasoning support through its Bayesian Network. AviatorDX will collect and reason upon observations from both subjective trainer assessments and through instrumentation to provide competency hypothesis validation and develop inferences with respect to demonstrated proficiency and the causes of observed performance deficits.TBD