The goal of this project is to address the knowledge gap surrounding human variability and transform this knowledge to fundamentally change human-system design from mitigating against to human variability to embracing and exploiting this variability. We aim to develop fundamental theories and models of human variability and uniqueness. Experimental efforts will focus on integrating three fundamental concepts:
- A multi-aspect approach that enables the interpretation of individuals “in context” (for discussion, see Oie et al 2012). To accomplish this, we aim to analyse and model integrated aspects of organization including: intended and unintended behaviors, physiology, subjective experiences, social interactions, environmental factors, and tasking.
- Sampling over long time scales and at high sampling rates. Long term high-resolution sampling is required to investigate the interaction between time-scales and to observe phenomena that occur rarely and over longer timescales (e.g. seasonal effects).
- Developing databases across a large population of people with adequate sub-sampling of Soldiers to enable the interpretation of individual differences and to more completely elaborate the breath of human variability.
To enable these experimental efforts, we will leverage a broad range of state-of-the-art technologies and tools for real-time sensing, computation and database needs, user modelling, and human performance prediction. Part of our effort will also seek to advance tools specific to long-term, continuous, multi-aspect assessment of humans with a focus on bridging gaps in current technology.
The Human Variability Project POC: Justin Brooks
UMBC Professors from the Mobile, Pervasive, and Sensor Systems Laboratory and Covail: Analog & Digital Systems Research Laboratory are developing novel sensors and virtual reality environments to collect human behavioral and physiology data to study the variability of human states in immersive environments.
University of Massachusetts – Amherst researchers in the Sensors Research Group are developing infrastructure to securely transmit, analyze, store, and visualize human data from numerous sensors and numerous individuals in real time.
THVP researchers are interested in understanding human behavior within natural environments. With the incredible diversity of behaviors individuals exhibit in these settings, one of the key challenges is to develop better methods to illustrate behaviors in the world. This example illustrates one of the concepts Alfred Yu, working in Innovation Commons, developed to depict individual coffee usage. By using a biodetector on the espresso machine and an augmented reality display, the concept is to have real time data of coffee consumption floating above the espresso machine. This concept could be used to understand coffee habits as a function of multiple timescales and individuals as well as inform and potentially help individuals with their specific habits.
Examples of the tools that the project aims to advance include:
- Information architecture and control framework for obtaining and integrating information sources
- Novel electronics packaging with flexible power management and storage solutions for Soldier-borne components.
- Approaches and algorithms to assess and predict non-linear human states that vary on multiple time scales across training and operational environments.
- Transfer learning, collaborative filtering, and other techniques to leverage information about other individuals, sub-groups of individuals, and groups to improve prediction of an individual or group.
- Techniques and fusion algorithms to interpret and predict non-stationary, human actions and behaviors in complex, dynamic, artifact-rich environments.