Fuzzy Mediation in Shared Control and Online Learning

Fuzzy Mediation in Shared Control and Online Learning

Giovanni Vincenti (Towson University, USA) and Goran Trajkovski (Algoco eLearning, USA)
Copyright: © 2009 |Pages: 23
DOI: 10.4018/978-1-59904-576-4.ch016
OnDemand PDF Download:
No Current Special Offers


This chapter presents an innovative approach to the field of information fusion. Fuzzy mediation differentiates itself from other algorithms, as this approach is dynamic in nature. The experiments reported in this work analyze the interaction of two distinct controllers as they try to maneuver an artificial agent through a path. Fuzzy mediation functions as a fusion engine to integrate the two inputs to produce a single output. Results show that fuzzy mediation is a valid method to mediate between two distinct controllers. The work reported in this chapter lays the foundation for the creation of an effective tool that uses positive feedback systems instead of negative ones to train human and nonhuman agents in the performance of control tasks.
Chapter Preview


Technology plays a dominant role as each new product is designed and placed on the market. When a technological solution oversees the cooking of rice, for example, one may or may not require much improvement. But, when we implement solutions that control a passenger jetliner, we need to be sure that the concepts and solutions are well founded.

The fly-by-wire airplane idea is similar to the one of an automobile that operates with little input from the driver, or a joystick-operated heavy-duty machine that lifts heavy loads. What these three examples have in common is the existence of mediation between signals from the operators and the actual sequence of actions taken by the machine in response to the operator. Sometimes, the response of the machine depends on the operator’s input; there are, however, times when the controls are autonomous responses to situations at hand.

Who (and when) should have control over the machine: the human or the automatic operator? Should we share control, and if so, how? Norman (2005) points out that what we understand as shared control is not really shared. For example, in early stages of development, the machine used to supervise and control any part of the operations in fly-by-wire airplanes, and now they oversee operations and still control flight. But, when the conditions don’t meet the standards of operations, the pilot is left alone. Therefore, it is either the pilot controlling the plane on his/her own, or the control system. There is no interaction between the two.

We need a system that allows for greater interaction between the human operator and the digital one. There are examples (e.g., Caterpillar machines) that allow operators to perform certain tasks only through automated systems (Grenoble O’Malley, 2005). Also, the two major airliner producers, Airbus and Boeing, are gearing toward advanced fly-by-wire technologies (Wallace, 2000), as the auto industry is attempting to infuse automation concepts in their products (Norman, 2005). These important initiatives stress that automation is becoming a predominant part of our everyday life. It is important, though, to realize that there is a need of some type of balance between controllers, humans, computers, or hybrids.As the interaction between machines and digital controllers increases, we need to:

  • 1.

    Find a better way to mediate control in dual control systems, when two (or more) operators are controlling the same machine;

  • 2.

    Then, replace one of the human operators with a digital one and investigate the interaction between the two entities using a mediation system (such as the one introduced in this chapter). This can be done if we:

  • 3.

    Create a framework for using simulation and virtual reality to test-drive solutions; and

  • 4.

    Implement these systems as a part of the actual operations of the machines under scrutiny.

This chapter will review some of the current problems in this “new” Information Fusion problem area, which extends the classic, static and traditional approaches to add a dynamic component that we call Fuzzy Mediation.

This chapter is organized as follows. First, we give background information about Information Fusion, as well as other preliminary concepts relevant to the framework for Fuzzy Mediation. Then, we introduce our original concept of Fuzzy Mediation as a theoretical solution to shortcomings highlighted. Next, we give a detailed overview of our algorithm and a breakdown of the three functional units that create it, and report on our conceptual experiments. Then, we introduce our application of Fuzzy Mediation to a robotic line follower simulation and discuss other problem fields that may implement this algorithm. Finally, we conclude the chapter.



In this section, we overview preliminary concepts needed to follow the rest of the chapter.

Control is a concept that involves the interaction of multiple entities (by definition at least two). In a situation of control, one of the subjects is identified as the one who interacts directly with one object, that is, directing the object’s every move (WordNET, n.a.).

Complete Chapter List

Search this Book: