Research in this field actually bridges the domains of biomechanics, control theory, and human factors to advance tightly coupled cyber-human systems, such as wearable technology. These efforts rely on three related sub-problems:

  • Characterization of variability for relevant functional tasks in a natural environment
  • Development of estimation algorithms, relevant computational models, and performance metrics that are robust to uncertainties found in the natural environment
  • Development of decision making interfaces that are synergistic with the expected end user

It is possible to approach each of these problems independently. However, these problems are all linked in that they involve an understanding of the human in a natural environment, use measured data to infer an aspect of the human system, and then must be presented to a decision maker (human or robotic system). The design and validation of algorithms and models are dependent on the data collected during the development phase. Knowing the data must be presented to a nonexpert may affect the selection of performance metric and thus how the algorithms are designed. Research from the human-centered viewpoint for decision making involves the understanding of the human in his or her actual environment, including the quantification of task variability, external confounding stimuli, the formalization of relevant expert knowledge, and presentation of the data to the decision maker. Through this unified path, we approach a wide spectrum of applications related to health and performance monitoring and system design, including for the astronaut, soldier, clinician, and general consumer.