Animating computer agents

To design automation to be a team player, we need to make it observable. Prototype concepts and design principles are introduced to demonstrate how to represent agent activity. These representations are based on field observations and a cognitive work analysis of a startup in a power plant. Our concepts aim to help practitioners to understand the basis for what the automation has done, what it is doing, and what it may do next. Three representation concepts in particular are tailored to automation-specific problems: contribution, continuous modifier and triggering logic graphics. Contribution graphics are used to depict how a set of independent factors contribute to a result showing the summation inherent in many algorithms. Continuous modifier graphics show agent adaptations by depicting how a contributor is discounted or amplified in a context-sensitive way. Triggering logic graphics illustrate how discrete conditional logic can be depicted in relation to many frames of reference.