Fuzzy mediation as a dynamic method of information fusion and shared control
Loading...
Links to Files
Permanent Link
Author/Creator
Author/Creator ORCID
Date
2010-07-27
Type of Work
Department
Towson University. Department of Computer and Information Sciences
Program
Citation of Original Publication
Rights
Copyright protected, all rights reserved.
There are no restrictions on access to this document. An internet release form signed by the author to display this document online is on file with Towson University Special Collections and Archives.
There are no restrictions on access to this document. An internet release form signed by the author to display this document online is on file with Towson University Special Collections and Archives.
Abstract
Information fusion algorithms are designed to perform efficiently when dealing with multiple sensors and environments with a static set of integration rules. When placed in an environment with multiple controllers and the need for dynamic fusion, conventional algorithms do not extend to accomplish the task. Fuzzy Mediation is an innovative approach to creating an information fusion framework that can be used to mediate between an expert and a novice controller. With this research, we explore different aspects of Fuzzy Mediation. We start with the demonstration of the concept in conceptual as well as static environments, and then we move on to simulations that place an expert controller and a one based on artificially intelligent algorithms at the commands of an agent. Finally we explore the reactions of human controllers to the Fuzzy Mediation framework in a setting of learning a new motor task. The results show that Fuzzy Mediation is a successful extension of conventional information fusion techniques to accommodate for the dynamic interaction of two controllers directing an agent. We observed that the overall performance of the agent based on Fuzzy Mediation shows an improvement when compared to the performance of an agent directed solely by the novice controller. Moreover, this framework allows for a novice with no training to start interacting with the agent after observing very few examples. This situation would not be possible if an unsupervised novice controller directed an agent. When human controllers interact, Fuzzy Mediation shows its strengths as an index of how closely the two operators perform. The results obtained through this research show that Fuzzy Mediation is a very promising dynamic extension of information fusion techniques. As this is a new approach to shared control, it sets the perfect ground for further research. The possible applications range from the automotive and aeronautical fields to the evaluation of training effectiveness and performance.