Abstract
This paper describes why and how the concept of adaptive automation was invented, starting with static function allocation, followed by dynamic function allocation. Levels of automation concept is utilized in explaining what kind of cooperation may be possible between human and automation in information acquisition, information analysis, decision and action selection, and action implementation. Sharing and trading are distinguished explicitly to clarify the style of human-automation cooperation. Algorithms for implementing adaptive automation are categorized into three groups, and comparisons are made among them. Decision authority and trust-related issues are also discussed. Benefits and costs of adaptive automation are described for better understanding and for future research activities.
