A Biocybernetic System for Adaptive Automation
Mark W. Scerbo Frederick G. Freeman Peter J. Mikulka Old Dominion University
Adaptive automation refers to systems that can adjust their mode or level of operation dynamically. Unlike traditional forms of automation where the operator is responsible for initiating changes in the state of the system, in adaptive technology both the operator and the system can initiate changes ( Hancock & Chignell, 1987; Rouse, 1976; Scerbo, 1996). Parasuraman, Bahri, Deaton, Morrison, and Barnes ( 1992) argue that adaptive automation can create a tighter coupling between an operator's level of workload and the degree or mode of automation in the system.
Interest in adaptive automation is fueled by concerns over the difficulties operators have when working with complex systems with multiple modes of automation ( Woods, 1996). Wiener and Curry ( 1980) indicate that it is not uncommon for pilots to become confused trying to keep track of the numerous modes of operation in modern flight-management systems. Further, Wiener ( 1989) argues that automated systems may actually become a burden at the times when a reduction in workload is needed the most. On the other hand, Parasuraman and his colleagues have shown that continued reliance on automation can impair the ability to detect system failures (see Parasuraman, Mouloua, Molloy, & Hilburn, 1996). An adaptive system would circumvent many of these problems by modifying its level or mode of operation in response to changes in the demands placed on the operator.
One of the critical problems facing designers of adaptive systems centers around the mechanisms for monitoring changes in workload and switching