HUMAN-COMPUTER COOPERATIVE PROBELM SOLVING IN SUPERVISORY CONTROL
Patricia Marie Jones
In the supervisory control of a highly automated system, the human operator must manage information and activities in a complex, dynamic environment. This thesis explores the concept of a computer-based, "intelligent" operators associate system as a means of providing context-sensitive, real-time aiding support to the human supervisory controller. First, a theory of human-computer cooperative problem solving is proposed that prescribes the requirements for an effective operators associate. Based on the theory and the results of a case study, the design of an operators associate for satellite ground control is described. The resulting system, the Georgia Tech Mission Operations Cooperative Assistant (GT-MOCA), was implemented in the context of an interactive, real-time simulation of the Mission Operations Room environment at NASA Goddard Space Flight Center. GT-MOCA provides task management, information management, and system visualization support for the human operator. This thesis provides a detailed description of GT-MOCAs operator modeling methodology, knowledge structures, and user interface. The results of an experimental evaluation of GT-MOCA by Flight Operations Team analysts at NASA Goddard are given, and show that GT-MOCA successfully addressed many of the issues in providing real-time support for human operators in the Mission Operations Room.