Dynamic Models of Augmented Cognition

Abstract

With changing technology and improving understanding of human neural mechanisms, the point is rapidly approaching where human-computer interaction could integrate information based on real-time cognitive state detection to adapt input to the user. This will result in a novel closed-loop system around a human operator. The shift to interactions around a closed-loop has the potential to produce fundamental changes in system performance of even well-understood open-loop scenarios. This article presents modeling based on engineering control systems theory that offers insight into such closed-loop systems. The model shows how dynamic instability can result from introducing feedback within a system and provides some methods that can be applied to remove such instability and optimize performance. The authors also examine the robustness of the closed-loop system to (parametric) variations in the (model of the) human operator. The use of such models allows for a systematic approach to analysis. This opens the door to many issues for future research, including system efficiency, design and optimization, as well as suitability of systems to variations across both operators and tasks. Some of the implications of such models for the future operation of human-computer systems are discussed, especially with a view to future work.