Date/Time
Date(s) - March 19, 2019
11:45 am - 12:35 pm
Location
CSE 122
Categories
Nicholas J. Napoli, Ph.D.
Research Scientist
Link Lab, Department of Electrical & Computer Engineering, University of Virginia
Abstract: Predicting Human Performance and Systemic Physiology through Multi-Modal Sensing
Capturing, characterizing, and predicting how humans perform within systems (e.g., semi-autonomous vehicles, smart health care, bio-feedback virtual reality platforms) is essential to closing the gap between humans and technology to create a symbiotic relationship. However, human performance data causes many problems that lead to ineffective system frameworks because of subject-to-subject variability, limited experimental data, generic feature engineering, inflexible models, lack of model modularity for new sensors, and inappropriate metrics. My research seeks to advance human-centered autonomous systems through the perspective of human performance by leveraging systemic physiology, bio-signal processing, and predictive modeling, which allows us to map the state of the human to the state of the system.
In this talk, we will discuss how hypoxia cognitively impairs pilots and how to enhance semi-autonomous aeronautic systems by developing novel methods for detecting errors in human performance using physiological measures. The work presented will focus on three fundamental areas of human systems: 1) Preventing performance metrics from becoming misleading indicators over time; 2) Improving detection of the state of the human by leveraging our understanding of physiology to develop novel bio-signal processing methods; and 3) Developing predictive modeling structures to detect performance errors caused by hypoxia.