Date of Award
5-2012
Document Type
Thesis
Degree Name
Master of Science (MS)
Legacy Department
Computer Science
Committee Chair/Advisor
Babu, Sabarish V
Committee Member
Hodges , Larry F
Committee Member
Duchowski , Andrew T
Abstract
A multi-agent scenario generation framework is designed, implemented and evaluated in the context of a preventive medicine education virtual reality system with data collected from a sensor network at the University of Iowa Hospital. An agent in the framework is a virtual human that represents a healthcare worker. The agent is able to make certain decisions based on the information it gathers from its surroundings in the virtual environment. Distributed sensor networks are becoming very commonplace in public areas for public safety and surveillance purposes. The data collected from these sensors can be visualized in a multi-agent simulation. The various components of the framework include generation of unique agents from the sensor data and low level behaviors such as path determination, directional traffic flows, collision avoidance and overtaking. The framework also includes a facility to prevent foot slippage with detailed animations during the travel period of the agents. Preventive medicine education is the process of educating health care workers about procedures that could mitigate the spread of infections in a hospital. We built an application called the 5 Moments of Hand Hygiene that educates health care workers on the times they are supposed to wash their hands when dealing with a patient. The purpose of the application was to increase the compliance rates of this CDC mandated preventive measure in hospitals across the nation. A user study was performed with 18 nursing students and 5 full-time nurses at the Clemson University School of Nursing to test the usability of the application developed and the realism of the scenario generation framework. The results of the study suggest that the behaviors generated by the framework are realistic and believable enough for use in preventive medicine education applications.
Recommended Citation
Gupta, Manan, "Data Driven Approach to Multi-Agent Low Level Behavior Generation in Medical Simulations" (2012). All Theses. 1348.
https://tigerprints.clemson.edu/all_theses/1348