Context-aware cognitive agent architecture for ambient user interfaces

Abstract

An ambient user interface is a set of hidden intelligent interfaces that recognize user's presence and provides services to immediate needs. There are several research activities on user interfaces and interactions which combining VR/AR, ubiquitous computing/ambient interfaces, and artificial intelligence. However, real-time and intelligent responses of user interfaces are still challenging problems. In this paper, we introduce the design of Context-aware Cognitive Agent Architecture (CCAA) for real-time and intelligent responses of ambient user interfaces in ubiquitous virtual reality, and discuss possible scenarios for realizing ambient interfaces. CCAA applies a vertically layered two-pass agent architecture with three layers. The three layers are AR (augmented reality) layer, CA (context-aware) layer, and AI layer. The two passes interconnect the layers as an input or output. One of the passes of each layer is an input path from a lower layer or environmental sensors describing a situation. The other pass is an output path and deliveries a set of appropriated actions based on the understanding of the situation. This architecture enables users interact with ambient smart objects through an ambient user interface in various ways of intelligence by exploiting context and AI techniques. Based on the architecture, several possible scenarios about recognition problems and higher level intelligent services for ambient interaction are suggested. © 2009 Springer Berlin Heidelberg.